Proceedings of the
1988 EPA/APCA International
Symposium:
Measurement of
and Related Air Pollutants
-------
Proceedings of the 1988
EPA/APCA international symposium on
MEASUREMENT OF TOXIC
AND RELATED
AIR POLLUTANTS
Jointly sponsored by the
US Environmental Protection Agency's
Environmental Monitoring Systems Laboratory
and
APCA - The Association Dedicated to
Air Pollution Control and Waste Management
Research Triangle Park, North Carolina
May 1988
-------
APCA Publication VIP-10
EPA Report No. 600/9-88-015
PROCEEDINGS OF THE 1988
EPA/APCA INTERNATIONAL SYMPOSIUM
ON MEASUREMENT OF
TOXIC AND RELATED AIR POLLUTANTS
NOTICE
Any policy issues discussed do not necessarily reflect the views of
EPA. Mention of trade names or commercial products does not con-
stitute endorsement or recommendation for use.
Published By:
APCA
P,O. Box 2861
Pittsburgh, PA 15230
11
-------
PREFACE
A joint conference for the third straight year, cosponsored by
APCA's TP-6, TP-7, and ITF-2 technical committees and the Environ-
mental Monitoring Systems Laboratory of the U.S. Environmental
Protection Agency was held in Research Triangle Park, North Caro-
lina, May 2-4,1988. The technical program consisted of 140 presen-
tations, held in 13 separate sessions, on recent advances in the
measurement and monitoring of toxic and related pollutants found
in ambient and source atmospheres. Covering a wide range of meas-
urement topics and supported by 51 exhibitors of instrumentation
and consulting services, the symposium was enthusiastically received
by more than 675 attendees from the United States and other coun-
tries. This volume contains the papers presented. The keynote address
to the symposium is also included.
Measurement and monitoring research efforts are designed to antic-
ipate potential environmental problems, to support regulatory actions
by developing an in-depth understanding of the nature and processes
that impact health and the ecology, to provide innovative means of
monitoring compliance with regulations and to evaluate the effec-
tiveness of health and environmental protection efforts through the
monitoring of long-term trends. EPA's Environmental Monitoring
Systems Laboratory, Research Triangle Park, North Carolina, is
responsible for research and development of new methods, techniques,
and systems for detection, identification, and characterization of pol-
lutants in emission sources and in indoor and ambient environments;
implementation of a national quality assurance program for air pol-
lution measurement systems; and supplying of technical support to
Agency regulatory programs.
This conference, the eighth in a series arranged each year by
EPA/RTP, but the third as a jointly sponsored conference by EPA and
APCA, was arranged with the following primary objective; to pro-
vide a forum for the exchange of ideas on the recent advances for the
acceptably reliable and accurate measurement and monitoring of toxic
and related pollutants found in ambient and source atmospheres. The
growing number of responses to this symposium represents an
encouraging step in the enhancement of our current measurement
and monitoring capabilities.
R.K.M. Jayanty and
Seymour Bochheiser
Technical Program Chairmen
in
-------
SYMPOSIUM COMMITTEES
TECHNICAL PROGRAM COMMITTEE
Cochairmen
Seymour Hochheiser, U.S. EPA
R. K. M. Jayanty, Research Triangle Institute
Donald Adams R.S. Braman J.E. Knoll
J.D. Spengler Terry Biddleman Gary Hunt
Thomas Shen Franklin Smith D.A. Lane
W.J. Dunn Charles Lochmuller J. Lewtas
Cliff Davidson
APCA TP-6 AMBIENT MEASUREMENTS COMMITTEE
Douglas Lane, Chairman
Thompson Pace, First Vice Chairman
Fred Dowling, Secretary
APCA TP-7 SOURCE MEASUREMENTS COMMITTEE
Billy Mullins, Jr., Chairman
Mark Siegler, Vice Chairman/Secretary
APCA ITF-2 TOXICS AIR POLLUTANTS
INTERCOMMITTEE TASK FORCE
David Patrick, Chairman
Jitendra Shah, Secretary
GENERAL CONFERENCE COMMITTEE
Cochairmen
Gary Foley, U.S. EPA
Martin Rivers, APCA
RESEARCH TRIANGLE PARK CHAPTER OFFICERS
Charles Pratt, Chairman
Mike Berry, Vice Chairman
Rodney Gibson, Secretary
Karen Gschwandtner, Treasurer
SOUTH ATLANTIC SECTION OFFICERS
Elizabeth Barfield, Chairman
James Southerland, Vice Chairman
Michael Tanchuk, Secretary
John Cline, Treasurer
John Jaksch, Membership Chairman
IV
-------
CONTENTS
The Role of Ambient and Source
Measurement in Implementing the
National Air Toxics Program
Strategy
Gerald Etnison
MEASUREMENT OF SEMI-VOLATILE
ORGANIC POLLUTANTS - AMBIENT AIR
Session Chairman: Donald F. Adams
Measurement of Semi-Volatile
Organic Pollutants - Overview Donald F. Adams
Recent Advances in On-site Measure-
ment of PCBvs With a Portable Gas
Chromatograph
Identification of Semivolatile Organic
Compounds in Selected Air Sample
Extracts by Gas Chromatography/
Matrix Isolation Infrared
Spectrometry
Coupled Supercritical Fluid Extrac-
tion/Gas Chromatographic Analysis
of Trace Organics From Atmospheric
Samples
Photochemical Aging of Polycylclic
Aromatic Hydrocarbons Found in Jet
Engine Exhaust
Analytical and Sampling Methods of
the Non-Occupational Pesticide
Exposure Study (NOPES)
A Systematic Procedure for Chlor-
dane Identification in Air
Chlorinated Pesticides and Poly-
chlorinated Biphenyls in the
Atmosphere of the Canadian Arctic
Methoxylated Phenols as Candidate
Tracers for Atmospheric Wood Smoke
Pollution
Organics Deposition Monitoring: The
Ontario Experience
Toxic Chemicals in Canadian
Rainfall
An Update on Grab Sampling of
Volatile Organics (VOC'a) and Other
Toxic Gases
A. Linenberg
Jeffrey W. Childers 15
Steven B. Hawthorne 2
Michael R. Kulhman 27
J. P. Hpu 34
Herbert J.
Sckattenberg, III 42
G.W. Patton 51
Steven B. Hawthorne 57
D.B. Orr
63
William M.J, StrachanlZ
Joseph P. Krasnec 78
-------
MEASUREMENT OF INDOOR (EXPOSURE)
TOXIC AIR CONTAMINANTS
Session Chairman: J.D, Spengler, Harvard School of Public Health
Measurement and Evaluation of per-
sonal Exposure to Aerosols
A Preliminary Study to Characterize
Indoor Particles in Three Homes
Asbestos in Residential
Environments
Nitric and Nitrous Acids in Environ-
mental Tobacco Smoke
High-Flow Personal Air Sampling as
a Component of Total Human
Exposure
Sampling and Chemical Characteriza-
tion of Workplace Atmospheres Con-
taminated with Airborne Diesel
Exhaust
EPA's Indoor Air Quality Test House
Mothcake Studies
Preliminary Results of the Baltimore
TEAM Study: I. Goals and Study
Design
Preliminary Results of the Baltimore
TEAM Study III. Indoor and Outdoor
Canister Measurements
Preliminary Results of the Baltimore
TEAM Study II. Personal Air and
Breath Measurements
Investigation of Environmental
Tobacco Smoke for Particulate Phase
Marker Compounds Using Mul-
tidimensional Gas Chromatography
R.W. Wiener 84
Richard Kamens 89
R.L Perkins 98
D.J. Eatougk 104
T.J. Buckley 113
R, A. Jenkins 119
Russell K. Clayton 131
William C. Nelson 137
A. Manale
LA. Wallace
Stanley L.
Kopceynski
143
149
155
ACIDIC DEPOSITION
Session Chairman: Cliff Davidson, Carnegie Mellon University
Design of a Glass Impactor for an
Annular Denuder/Filter Pack System
Comparison of Methods for Monitor-
ing Dry Deposition Pollutants: Sum-
mer 1987 Study
P. Koutrakis 164
E. Hunter
Daughtrey, Jr. 170
VI
-------
Measurement of Atmospheric Aerosol
Acidity: Losses From Interactions of
Collected Particles
Introduction of a NO/NO2 Monitor for
the Sub-ppb Range
Dry Deposition of Ca and Mg to
Selected Trees in an Eastern Kansas
Oak-Hickory Forest During Sequen-
tial Synoptic Events
Aerosol Nitrogen Inputs to a
Tree/Grass Ecotone: Project Overview
Toxics in Fog and Their Potential
Environmental Problems
Organic Chemical Characterization of
Clouds in High Elevation Spruce-Fir
Forests at Mt. Mitchell, North
Carolina
Relative Importance of Dry, Wet, and
Cloud Capture Mechanisms for Acidic
Deposition
Source Signatures in Cloud Water
J.L. Slater
Werner Martin
176
182
Mark J. Thomas
Steven C. Mauch
Kumar Ganesan
Viney P. Aneja
V. K. Saxena
Ilhan Olmez
189
200
216
227
237
243
MEASUREMENT OF VOLATILE ORGANIC
POLLUTANTS - AMBIENT AIR
Session Chairman: Charles H. Lochmuller, Duke University
Application of Cryogenic Trapping
and Two-Dimensional Gas Chro-
matography for the Measurement of
Atmospheric Oxygenated
Hydrocarbons
Charng-Ching Lin 251
Comparison of a Cryogenic Precon-
centration Technique and Direct
Injection for the Gas Chromato-
graphic Analysis of Low ppb
(NMOL/MOL) Gas Standards of Toxic
Organic Compounds
Automated Analysis of Multicompo-
nent Compressed Gas Mixtures Con-
taining Parts Per Billion
Concentration of Toxic Organic
Compounds
George C. Rhoderick 259
Gary B, Howe
265
Vll
-------
A New Method for Analysis of VOCs
in Kanawha Valley Ambient Air
Evaluation of a Tekmar 5000 Ther-
mal Desorber for Use with a Hewlett-
Packard GC/MSD for the Analysis of
Volatile Organic Compounds Col-
lected on Tenax from Ambient Air
Application of an lon-Trap/Mass Spec-
trometer to the Analysis of Ambient
Organic Emissions
Sampling and Analysis of Toxic Vola-
tile Organic Pollutants in Ambient
Air Using an Automatic Canister-
Based Sampler
Air Toxics Interface and Analytical
Systems for Ambient Air Samples
The Use of Tedlar Bags for
Integrated Gaseous Toxic Sampling:
The San Francisco Bay Area
Experience
Comparison of Evacuated Flasks and
Tenax for Detection of Selected Com-
pounds Under Controlled Condition
Identifying the Composition of
Source-Related Groups of Volatile
Organics in the Ambient and Indoor
Air
A Second Generation Network Design
for VOC Field Sampling Using
Whole-Air Samplers
Analysis of Toxic Organic Vapors in
Air Using a Portable Photoionization
Gas Chromatograph
The Use of a Photoionization Based
Gas Chromatograph for the Analysis
of Field Samples
Artifacts in the Sampling of Ambient
Organic Aerosols
Potential Atmospheric Fate of VOC's
Analyzed in the 1987-1988 Denver
IEMP
B. DasSarma
277
Rebecca A. LaRue 285
Robert G. Orth 291
Clyde W. Sweet 299
Dave-Paul Dayton 305
D. A. Levaggi
313
Richard W. Tripp 324
William A. McClenny 331
Glen A. Marotz 341
Richard E. Berkley 352
Michael Duffy 358
Kochy Fung 369
Gordon E. Pierce 377
Vlll
-------
MEASUREMENT OF HAZARDOUS WASTE EMISSIONS
Session Chairman: Thomas T. Shen, New York State Department
of Environmental Conservation
Overview of Applicable Emission
Measurement Technologies for the
Measurement of Volatile Hazardous
Waste Emissions
J.A. Clark
383
Planning Air Monitoring for VOCs at
Waste Sites
Preliminary Evaluation of Test
Methods for Volatile Organics in
Hazardous Waste
Air Emissions from Hazardous Waste
Stabilization
Integrating Sampler for Hazardous
Pollutants in Liquid Streams
Validation of Analytical Methods for
Determining Total Chlorine in Used
Oils and Oil Fuels
Evaluation of Performance of Carbon
Monoxide Monitors on Hazardous
Waste Incinerators
Laboratory Evaluation of a Test
Method for Measuring Emissions of
Selected Toxic Metals from the
Incineration of Hazardous Materials
Emissions from Hazardous Waste
Incinerators; Design or Operating
Problems
Multimedia, Multipollutant Field
Study to Establish Levels of Toxic
Contaminants in Air, Soil, Sediment,
Water and Agricultural Products
from a Model Municipal Waste
Combustor
Assessment of 2,3,7,8-TCDD Emis-
sions from Waste Disposal Sites
The Importance of Proper Site
Characterization of the Contaminant
Pathway
Michael J. Barboza 399
Sam B. Balik
406
Paul R. de Percin 413
R. G. Merrill, Jr. 418
A. Gaskill, Jr.
426
Larry Edwards 432
Nancy F. Cole 441
Joseph J. Santoleri 447
L. Fradkin
461
Seong T. Hwang 470
C.E. Schmidt
486
IX
-------
SOURCE MONITORING
Session Chairman: J.E. Knoll, U.S. Environmental
Protection Agency
Field Test Evaluation of a Methodol-
ogy for Measuring Emissions of
Selected Toxic Metals from Station-
ary Sources
GlenD. Osmond 497
In-Stack PM10 Sampling Methods: A
Review of Basic Requirements
Development of Methodology to Meas-
ure Condensable Emissions from Sta-
tionary Sources
Particulate Matter-Organic Com-
pound Interactions, Municipal
Incinerator Fly Ash Studies
Measurement of Ethylene Oxide
Emissions from Hospital Sterilizers
Evaluation of Sampling Methods for
Measuring Ethylene Oxide Emissions
from Sterilization Chambers and Con-
trol Units and Determining Control
Unit Efficiency
Feasibility Study on Real Time Meas-
urement of Toxic Incinerator Emis-
sions with a Trace Atmospheric Gas
Analyzer
Studies on the VOC Analytical
Method by the Use of a TOC
Analyzer
William E. Farthing 503
J.D. McCain
510
R.G. Merrill, Jr. 517
P.T. Leclair
524
J.L. Steger
L.E. Slivon
M.R. Peterson
530
536
541
CHEMOMETRICS AND ENVIRONMENTAL
DATA ANALYSIS
Session Chairman: W.J. Dunn, University of Illinois at Chicago
The Use of Fractal Dimension to
Characterize Individual Airborne
Particles
Statistical Detection of Changes in
Ambient Levels of Toxic Air
Pollutants
Industrial Toxic Gas Storage Facility-
Dispersion Study
Philip K. Hopke 548
Mithra Moezzi 556
Douglas R. Murray 569
-------
Statistical Properties of Hourly Con-
centrations of Volatile Organic Com-
pounds at Baton Rouge, Louisiana
Alison K. Pollack 575
MEASUREMENT OF TCDDs/TCDFs IN AMBIENT AIR
Session Chairman: Gary T. Hunt, ERT, Incorporated
Evaluation of High-Volume Sampling
Techniques for the Determination of
CDD/CDF in Ambient Air
Evaluation of the Collection Effi-
ciency of a High Volume Sampler
Fitted with an Organic Sampling
Module for Collection of Specific Poly-
halogenated Dibenzodioxin and
Dibenzofuran Isomers Present in
Ambient Air
C. Tashiro
590
T. O. Tiernan
Monitoring Ambient Air for Dioxins Bill J. Fairless
Determination of Polychlorinated
Dibenzo-P-dioxins and Dibenzofurans
in Stack Gas Emissions and Ambient
Air
596
602
Robert L. Harless 613
Intercomparison Study of Ambient
Air Dioxin/Furan Sampling and Ana-
lytical Methods
Background Environmental Concen-
trations of Dioxins and Furans
Congener Profiles of Polychlorinated
Dibenzo-P-dioxins and Dibenzofurans
in Atmospheric Samples
T. Dann
Brian Eitzer
621
629
Jean M. Czuczwa 634
GENERAL
Session Chairman: R. S. Braman, University of South Florida
Dielectrophoresis of Chloroplasts-A
New Technique in Biomonitoring of
Low Levels of SOZ
Some Problems and Considerations
Related to Airborne Asbestos Sam-
pling in the Outdoor Environment
The Preparation of Summa Canister
Performance Samples and Their Sub-
sequent Analysis by the TAGA
6000E MS/MS
Adeel Ahmed 640
David R. Suder 645
Rita M. Harrell 655
XI
-------
The Impact of Residential Wood Com-
bustion on Ambient Wintertime Car-
bon Monoxide Concentrations in
Residential Areas in Six North-
western Cities
Utilization of Carbon-Based Adsor-
bents for Monitoring Adsorbates in
Various Sampling Modes of
Operation
Design of a Sampler for Peroxyacetyl
Nitrate Monitoring
Performance Evaluation of the Har-
vard/EPA Annular Denuder Systems
Under Simulated Atmospheres
Operation of Annular Denuder Sys-
tems for Atmospheric Acidity Meas-
urements in Bilthoven, The
Netherlands
Development of a Sampling Proce-
dure for Large Nitrogenous Particles:
Preliminary Results
EPA's Indoor Air Quality Test House
2. Kerosene Heater Studies
Design of a Self-Administered Per-
sonal Daily Activity Questionnaire
for Evaluating Exposure to Combus-
tion Products
A Low Cost Data Acquisition System
for Residential Combustion Spillage
Monitoring
Background and Status of Computer-
ized Systems for SARA Title III
Emergency Planning for Accidental
Chemical Releases
James E, Houck 664
William R. Betz 670
Kochy Fung
679
Michael Brauer 685
Jed Waldman
691
Dennis D. Lane 699
Merrill D. Jackson 715
N. C. G. Freeman 720
Mark Lawton
727
Jane C. Bare
733
MEASUREMENT OF VOLATILE ORGANICS
AMBIENT AIR
Session Chairman: D. A. Lane
Sampling Gaseous compounds in
Environmental Tobacco Smoke
D.J. Eatough
739
Xll
-------
Results from the Environmental
Response Team's Preliminary Evalu-
ation of a Direct Air Sampling Mass
Spectrometer (the Bruker MM-1)
Review of Gas Phase Retention Vol-
ume Behavior or Organic Compounds
on Tenax-GC and Other Sorbent
Materials
Chamber Studies Characterizing
Organic Emissions from Unvented
Kerosene Space Heaters: Phase II
Results from the Environmental
Response Team's Evaluation of the
TAGA 6000E Direct Air Sampling
Mass Spectrometer/Mass Spectrometer
A Study of Products From the Photo-
oxidation of Toluene Using MS/MS
Analysis
Field Comparison Study of the Com-
bustion Engineering 8202A and
Integrated Grab Sample/
Preconcentration Direct Flame loni-
zation Detection for Ambient Meas-
urements of Non-Methane
Hydrocarbons
Robert E. Hague 750
James F. Pankow 765
Patricia M. Boone 769
Thomas H. Pritchett 775
T.E. Kleindienst
787
Joel C. Craig
793
INTEGRATED AIR CANCER PROJECT STUDY
Session Chairman: J. Lewtax, U.S. Environmental
Protection Agency
The Integrated Air Cancer Project:
Overview and Boise Survey Results
Influence of Residential Wood Com-
bustion Emissions on Indoor Air
Quality of Boise, Idaho Residences
Distribution of Volatile Organic
Hydrocarbons and Aldehydes During
the IACP Boise, Idaho Residential
Study
Semivolatile and Condensible
Extractable Organic Materials Distri-
bution in Ambient Air and Wood-
stove Emissions
Larry T. Cupitt 799
V. Ross Highsmith 804
Roy Zweidinger 814
R.G. Merrill, Jr. 821
Kill
-------
GC/MS Analysis of Woodstove Emis-
sions and Ambient Samples From a
Wood Smoke Impacted Area
Effects of Operating Variables on
Emissions from Woodstoves
Impact of Residential Wood Combus-
tion and Automotive Emissions on
the Boise, Idaho, Airshed
What Should We Measure? Aerosol
Data: Past and Future
Sources of Fine Particle Organic Mat-
ter in Boise
R.S. Steiber
828
Robert C. McCrillis 835
V. Ross Highsmith 841
L. A. Currie
Charles W. Lewis
Annular Denuder Results from Boise,
ID R. K. Stevens
Mutagenicity of Organics Associated
with PM2.S and PM10 HiVol Particles
from a Wood Smoke Impacted
Residential Area
Transformation of Boise Sources: The
Production and Distribution of Muta-
genic Compounds in Wood Smoke
and Auto Exhaust
Final Design and Field Evaluation of
the High Volume PM2.S Virtual
Impactor
853
864
870
R. Watts
L.T. Cupitt
Robert M. Burton
879
885
890
ENVIRONMENTAL QUALITY ASSURANCE
Session Chairman: Franklin Smith, Research Triangle Institute
Quality Assurance Plan Used at the
Love Canal Emergency Declaration
Area Indoor Analyses by a TAGA
6000E Mass Spectrometer/Mass
Spectrometer
Thomas H. Pritchett 896
Quality Assurance for Personal
Exposure Monitoring - An Update
Considerations in the Design of Air
Toxics Monitoring Programs at
Superfund Sites
D. J. Smith
914
Richard Crume 922
xiv
-------
Statistical Analysis of GC/MS Perfor-
mance Audit Data Raymond C. Rhodes 932
Quality Assurance for Measurement
by EPA Method 5G for Wood Heater
Certification Testing Glenn D. Rives 939
An Alternative Standardization
Method for the Analysis of Gaseous
Organic Compounds Thomas Berstiel 945
Interpretation of Field Performance
Audit Data in Woodstove Emission
Measurement Programs Joseph D. Evans 953
Index 963
XV
-------
The Role Of Ambient And Source Measurement In Implementing The
National Air Toxics Program Strategy
Gerald Emison, Director
Office of Air Quality Planning and Standards
Summary
Mr. Emi son began by expressing his pleasure in speaking at the 3rd
Annual EPA/APCA Symposium on Measurement of Toxic and Related Air Pollutants.
He commented on the growth of the Symposi urn - the number of attendees has
more than tripled in 3 years. He was impressed with the diversity of
topics to be discussed - measurenent systens and approaches, source
monitoring, the Integrated Air Cancer Project Study, quality assurance,
indoor air, and the statistical analysis of environmental data - and with
all the professionals in the air pollution control community attending
the symposi urn. Participants came from academia, the private sector,
various State, local and Federal agencies, and from Canada and Germany as
wel 1 .
Mr. Emison began by emphasizing the importance of the Conference. He
asked two questions: "Why should a group of researchers be interested in
hearing from me? Why, as a representative of the Air Program, should I
be here talking to you?" His answer: "We need each other," One of the
most dynamic challenges facing EPA and, in fact, facing all of us
professionals in the air pollution control community, is that of
air toxics management.
The National Air Toxics Program Strategy requires a partnership
sffort among Federal, State and private sectors. While not playing down
the Federal/State role, he highlighted the importance of the private
sector. He cited the private sector's role in providing air toxics
information (such as through SARA Title III), accident prevention, the
research and development of ambient and stack monitoring instrumentation,
and participation in both the National Air Pollution Technical Advisory
Committee Review and the Science Advisory Board.
At the national level , he stated that the problem is regulated by
9'i of the Federal authorities - NSPS, RCRA, TSCA, Mobile Sources and
Action 112 of the Clean Air Act (CAA). The overall strategy is managed,
technical and financi al support i s given to State and local agencies, and
research and development activities are carried out to support the national
strategy. The total ORD air quality research budget was $65.7 million in
£' 88, and air toxics, at $22.3 million, was the largest component. The air
toxics budget includes scientific assessment; monitoring and quality assurance;
health effects; control technology; and the characterization, transport
afid fate of air toxics pollutants. Of these, the monitoring and qual ity
assurance component is the largest, with $7.6 million.
-------
The State and local agencies play a major role in evaluating and
controlling point sources not regulated at the Federal level, and in
assessing toxic exposures from multi-pollutant sources. He stated that
the Federal government is working with the State and local agencies to
enhance their own program capability, not only to help them accomplish
these objectives but also to help them implement their own statutes.
Mr. Emison then discussed the Federal Pollutant Assessment Program,
which is designed to evaluate and to decide the need for regulating specific
chemicals and groupings of chemicals, or sources that emit mixtures of
chemicals (coke ovens, municiple waste combustors, coal and oil combustion).
Both health effects and exposure levels are examined to determine if
regulation is needed.
The Federal Regulatory Program was discussed in the context of the
vinyl chloride decision, a unanimous opinion of the Washington, DC Circuit
Court in July 1987. The Agency's position on this case was that Section
112 did not require a zero level of emissions for a carcinogen presumed
to pose a cancer risk at any level of exposure. The Court suggested that
the Agency can consider cost and feasibility in setting a standard to
protect public health. The court decision envisions a two step process
in standard setting: (1) EPA is required to determine a "safe" or "acceptable
level of risk without considerations of cost or feasibility, and (2) the
Administrator provides an "ample margin of safety", which can consider
cost and feasibility. What is EPA doing as a result of this? EPA is now
looking at ways to define "acceptable risk". "Acceptable" doesn't have
to be zero, but there is little guidance on what the Court considers
"safe".
The Federal Regulatory Program under the Clean Air Act includes
Section 112, the hazardous air pollutant provision, as well as other
parts of the CAA, which can effect reductions of toxic air emissions.
The recently promulgated PM^Q NAAQS will require reductions in particulate
emissions, which should cause reductions in air toxics exposures from
constituents such as metals. Similarly, in order to attain the ozone
standard, volatile organic compound emissions will be reduced, which will
also lessen exposure to toxic organic compounds. Through these programs,
significant indirect control of toxic emissions is expected. The current
program, under the CAA, also includes NSPS, NESHAPS - Section 112, and
Mobile Sources. Mobile sources involves control of diesel particulate,
onboard controls, the catalyst program, methanol regulations, fuel additive
and composition regulations, and antitampering and misfueling requirements.
The current authority to control air toxic emissions also includes
other authorities: RCRA, TSCA, FIFRA, CERCLA and CWA. In summary, the
Federal level involvement is extremely broad based.
The State and local programs have four major themes: (1) accept
NESHAP enforcement delegation, (2) address high risk point sources - not
regulated by Federal programs, (3) address high risk urban problems,
and (4) enhance State and local ability to identify their own specific
problem areas and to implement their own legislation and regulations.
-------
The air toxics problem is tough and complicated. It is a very extensive
and technically uncertain problem. There are serious effects, and the
public is interested. The sources are not principally large industries,
but rather area sources and consumer products, and the regulatory authority
is diffuse.
Mr. Emison discussed financial and technical support. Financial aspects
include $11 million for State and local agencies, while technical support
includes the Control Technology Center, the Exposure Center, air toxics
research and development, guidance and monitoring strategy, and workshops
and training.
His address emphasized the Urban Air Toxics Program (UATP), which
assists in characterizing the nature and magnitude of the air toxics
problem through ambient monitoring and emission inventories. Technical
support under UATP includes inventory and monitoring guidance.
For the future, Mr. Emison identified three major needs: (1) information
on the origin and individual species of toxic, mutagenic or carcinogenic
pollutants present in ambient air, (2) understanding of the changes in
chemistry and mutagencity which occur between emission of a pollutant and
a person's exposure to it, and (3) a need for new monitoring methods for peak
measurement and pattern recognition of VOCs, polar organics (such as
ethyl ene oxide), semivolatile organic chemicals, biomarkers for assessing
health risks of human exposure to toxic air pollutants, and remote sensing.
In order to accomplish these goals, EPA needs to develop and maintain
research capabilities through the Toxic Air Pollutant Program.
In closing, Mr. Emison stated that the air toxics problem is a
tough one, but we are all moving ahead. Almost without exception, we have
the opportunity to set our direction in a way that allows each of us to
contribute his or her own unique strengths. The State and local governments
are doing what they are best suited for, and the Federal government is
doing what it is best suited for. Control of air toxics is a problem
that is amenable to solution through the efforts of professionals dedicated
to protecting the public health, collaborating at all levels of government.
Mr. Emison ended by saying that he looked forward to working with everyone
involved in this problem, with the expectation that jointly we will solve
]t, and the national air pollution program will continue to protect the
health of all the people.
-------
MEASUREMENT OF SEMI-VOLATILE
ORGANIC POLLUTANTS - OVERVIEW
Donald F. Adams, Consultant
SW 600 Crestview #2
Pullman, WA 99163
INTRODUCTION. Semi-volatile organic compounds (SVOCs) may be
present in the atmosphere as vapors and adsorbed on suspended particulate
matter. Junge (1) stated that in clean air and 25*C most organic
compounds with vapor pressures below 10 torr (mm Hgi exist in the
vapor phase and compounds with vapor pressures above 10 torr exist in
the particle phajse. The SVOCs generally fall in the vapor pressure range
of 10 to 10 torr. Lewis (2) later suggested the range for SVOCs
should include those compounds having vapor pressures between 0.1 and
10~ torr.
Broadly speaking, SVOCs encompass those compounds which are too
volatile to be collected and completely retained by air filtration
sampling. Even the widely used two-stage filtration/back-up adsorption
sampling methods are not entirely without question because of the
possibility of incomplete SVOC retention and artifact formation during the
sampling, storage, and desorption processes.
A wide variety of SVOCs are found in the atmosphere including
polychlorinated biphenyls, dioxins and furans; polycyclic aromatic
hydrocarbons; agricultural and household pesticides; and herbicides.
These compounds are produced by many sources such as combustion processes
including internal combustion engines, municipal and industrial waste
incineration, electric power and steam generation, household wood and
kerosene heaters, and tobacco smoke; commercial chemical synthesis
processes; toxic chemical spills; hazardous waste storage and cleanup;
agricultural, commercial and residential use of commercial chemical
products; etc. Due to their disperse sources, these chemicals are widely
found in ambient and indoor air.
-------
Historically, SVOC sampling methods relied solely on filtration for
collection and pre-concentration. However, it is now recognized that an
appropriate SVOC sampling system must involve at least a combination of
particle filtration and some type of back-up vapor phase collector.
DISCUSSION. Sample pre-concentration from the bulk air matrix is
required to obtain an adequate mass of material for the detection and
quantification of the low concentrations of SVOCs usually present in
ambient air even when using the most senstive analytical techniques.
Therefore, it is essential to consider problems associated with pre-
concentration techniques before any sampling program is undertaken.
We are primarily concerned with the partitioning in the sampling
system and assurance that a total SVOC sample is collected. Although
partitioning of the SVOCs from the collected particles to the gas phase
vor from the gas phase to the particles) begins at the source(s) and
continues in a dynamic manner throughout the transit time and distance of
the mixture to the sampler, we can only attempt to understand and account
for these processes which may occur in the receptor sampler.
At the sampler, partitioning of the SVOCs from the filter to the gas
phase continues while the particulate matter is collecting on the filter.
The extent of the loss or gain of SVOCs from filter-collected particles
will depend on their vapor pressures; the associated particle composition,
loading, and size distribution; and the temperature and humidity. It is
•also possible that some collected particles may adsorb some vapor from the
air samples.
These losses or gains vary with the ambient temperature and the sam-
pling time following an elevated fumigation episode during any given eight-
or 24-hour sampling period because of the continuous and variable loss of
the collected SVOCs to the sample air stream during the remainder of the
sampling period. The temperature effect is also important, since the
vapor pressure of SVOCs nearly doubles with each 5*C increase. Whether
sample loss or gain predominates depends on how and when these variables
change during a sampling period as well as the variations in airborne SVOC
vapor concentrations during each sampling period.
Even with a back-up vapor collector, the derived analytical data do
n°t realistically reflect the airborne distributions, although tandem
sample collection may efficiently retain both particle and vapor phases.
Only the total airborne concentrations of SVOCs can be determined with
two-stage particle/vapor collection SVOC air sampling methods because they
8eparately collect phase-distributed SVOCs.
Artifact formation must also be considered. Artifact varies with
condensation or vaporization during the sampling process and also via
possible reaction with ozone or other reactive species.
For analysis, however, the filter catch should be combined with the
back-up collection. Although separate analyses of each of the two-stage
c°llections will provide distribution information between the filter and
the back-up collector, these results can not be interpreted to reflect the
distribution between the gas and aerosol phases existing in the atmosphere
immediately prior to collection.
-------
The widely-used, two-stage SVOC sampling methods may not adequately
distinguish between phases due to probable disruption of the vapor/
particle equilibria during filtration and the inability to collect
separately the vapor phase SVOCs present in the air sample prior to
filtration.
Insight into this unresolved problem may be provided by using a
three-stage sampler using a first-stage vapor phase denuder followed by
the usual particle filter and a back-up vapor collector (3-6). The
analytical data may be interpreted in terms of the amount of SVOCs in the
vapor phase (stage 1), SVOCs associated with the particulate matter (stage
2), and SVOCs volatilized from the particulate matter during sampling
(stage 3).
Further work is required to establish the vapor phase collection
efficiency of denuders for the wide range of SVOCs to establish the best
denuneder coating for general application.
REFERENCES
1. C. E. Junge. "Basic Considerations about Trace Constituents in the
Atmosphere as Related to the Fates of Global Pollutants." Fate of
Pollutants in the Air and Water Environments, Part I, pp. 7-25.
I. H. Suffit, Ed. John Wiley and Sons~New York. 1977.
2. R. G. Lewis. "Problems Associated with Sampling for Semi-volatile
Organic Chemicals in Air." Proceedings of the 1986 EPA/APCA
Symposium on Measurement of Toxic Air Pollutants, pp. 134-145.
APCA Publication VlP-7. Air Pollution Control Association.
Pittsburgh, PA. 1986.
3. N.
-------
RECENT ADVANCES IN ON-SITE MEASUREMENT OF PCB'S WITH A
PORTABLE GAS CHROHATOGRAPH
D>~. A. Linenberg
President
Sentex Sensing Technology, Inc.
Ridgefield, New Jersey
INTRODUCTION
The EPA recently documented specific sampling and
analysis methods to determine compliance with the USEPA
National Spill Cleanup Policy. (1) (2). Gas Chromatography
with an electron capture detector was recommended as the
Primary method of analysis. The advantages of using field
Portable GC's for real-time analysis, thus eliminating
delays in taking remedial action and the need for return
visits to the spill site, were also highlighted.
This paper describes hou the latest developments in
a field portable GC simplifies accurate on site analysis
when used with established EPA procedures for sampling and
sample preparation.
On site determination of PCB's require certain pro-
cedures and instrument features which can be summarized as
A sample preparation procedure;
The sampling method developed by Dr. Spittler (5)
was used and found satisfactory for the test
sample preparation.
An instrument properly equipped for conducting the
analysis;
The system used for the program was the Scent ograph
Portable Gas Chroma t ograph equipped with the
following features;
1. On column heated injection port
-------
2. Temperature controlled/programmable column oven
3. 10 meter capillary column (Ni,.5MM I.D. )
4. Electron Capture Detector
5. Detachable lap-top computer (For running the G.C.
display evaluation and storage of data)
C. A procedure for on site interpretation and evalua-
tion of the results as follows:
1. Establishment of calibration chroroatograms for
various PCB standards;
2. Recording of all retention times of various
peaks;
3, Calculating the area of the peaks;
4. Performing the analysis chromatograms;
5. Comparison of retention times of the analysis
chromatograms with those of various calibratants;
6. Comparison of peak areas of analysis chromatograms
with the peak areas of calibration chromatograms
and calculation of concentration levels.
Conducting these steps, on site* is the most complicated
part and, in the past, has detered operators from performing
on site analysis.
The Scentograph, however, equipped with the lap-top
computer, includes a software package specially designed for
this purpose. The sequence of operation is as follows:
1) All calibration chromatograms of known PCB standards are
conducted prior to or after on site analysis. The chro—
matograms, retention times and peak areas are stored in
the Scentograph memory.
2) Analysis is performed in the same manner and the results
are also stored in the instrument's memory.
3) Calibration chromatograms can now be recalled from memory
and displayed against analysis results. The computer
matches retention times to give the best PCB recognition.
In addition, the computer.will calculate the total area of
the analysis peaks compared with the calibration. Accord-
ingly, the concentration levels can be easily evaluated.
D. Recordkeeping:
All results are stored In the Scentograph's memory for
future documentation. The Scentograph can store ZOO to
500O chromatograms on a single floppy disk which also
contains the entire details of the analysis, such as
dates, times, retention times, area counts, and the
actual chromatograms.
8
-------
EXPERIMENTAL;
Calibration Preparation:
With the Scentograph parameters set as shown in Figure 1,
PCB standards are injected into the instrument and the
resultant chromatograms stored on disc in the calibration
file.
On-Site Sampling:
Standard EPA procedures For hexagonal grid sampling (3) are
used.
Sample Preparation:
As developed by USEPA Region I (4) field samples are prepared
by Placing 40O rag. of soil into a 3 cc septum vial. 1OO
nucroliters of water was added to the soil; 4OO microliters
of methanol and 5OO microliters of hexane were also added.
Agitation was performed by shaking for 20 seconds. A sample
was taken from the top layer in the vial with a 10 microliter
syringe and 2 microliter was injected into the Scentograph.
The Scentograph was set with the same parameters used for
analysis of the standard calibration samples.
RESULTS AND DISCUSSION;
The chromatograms for calibration standards of AROCLOR's
1016,1242,1248 and 126O at 50 ppm are shown in Figures 2-6.
n each case the total chromatograms was completed in less
than 15 minutes at 16O deg. c.
When an unknown sample was tested on—site the resultant
chromatogram was compared with the calibration standards
hrough the simultaneous display, on the computer screen,
t both analysis and calibration chromatograms until a
fingerprint match is obtained. Figure 6 shows no match,
therefore sample is not Aroclor 126O. Figure 7 shows a
watch, to Aroclor 1242. The Scentograph, automatically,
integrates the total chromatogram of the standard sample
and asigns it a value of 1OO percent. Accordingly, 1OO
Percent becomes equivalent to 5O ppm, if a 5O ppm calibra-
ion standard has been used. The complete area of an anal-
ysis chromatogram is integrated and given a percent value
compared to the calibration chromatogram calibrated as 1OO
Percent. Thus a 12O percent value will represent a 6O ppm
concentration level or an 8O percent value which repre-
sents a 40 ppm sample. Using this technique it is extremely
to identify and quantify the field sample.
9
-------
CONCLUSION;
Recent developments in portable GC's has significantly
simplified. PCB analysis in the field. Operating conditions
and calibration/recognition data may be preset. Identifica-
tion of PCB is by direct onscreen comparison of fingerprint
chromatograms. Quantification is by comparison of total
integrated areas. Recordkeeping is automatic and data is
stored in a report ready format.
REFERENCES;
1. USEPA, Verification of PCB Cleanup by Sampling and
Analysis* Miduest Research Institute Report for Office
of Toxic Substances, EPA-56O/5-85-O26, August 1985
2. Environmental Protection Agency, Polychlorinated
Biphenyls Spill Cleanup Policy (4O CFR 76)
Federal Register Vol. 52, No. 63, April 2, 1987
(10688 - 10710)
3. USEPA, Field Manual for Grid Sampling of PCB Spill
Sites to Verify Cleanup, Midwest Research Institute
Report for Office of Toxic Substances, EPA 56O/5-86-O17
May 1986
4. Dr. T.M. Spittler, (EPA Region I), PCB Soil Screening
Procedure, Lexington, Mass.
10
-------
OPERA!ING PARAHETERS
1-
-Co.liNation Sample NaMe (enter
* 1 *!• • •
- '*a
|>
8
18
li
12
13
14
15
Oven
han Duration
Analysis Per Gal lotion
uto Analysis Duration
BacHlush Option
(enter
l&-
CohiHfi
Col'tsn Pressure
* of Calibation Peaks
Peak Nunber i
Substance Na«e
Coriceritration Range
Calibration Cone
Peak *larn values ( 0 • 99,99
upload Scentogpaph paraHeters
up to 8 letters)
( 1 - 380 sec)
( 0,1 - 4,0 sec)
( 0,1 - 4,0 sec)
( 18 - 999 sec)
( 30 - 180 °C )
10, 15, 20,38 win)
( i - 99 )
(0-120 Nin)
( flroff. Iron)
ECD,3-TCM-PID>
up to 8 letters)
(5-40)
(1-16)
AROC1260
i
0,1
0,1
30
160
15
99
MANUAL
BCKFL OFF
E.C.D,
CAP 10' i
20
16
(enter
( PPH :
( 99,99
UP to 8 letters) ,, PCB i
0, PPBrl),, PPM
PPM, 9999 ppb) ,, 50,00
PPM / 9999 ppb) ,, 1
Enter- field to be. UPDATED? I
Figure
12:85:46
: AROC1016
Figure 2
Press (ENTER)? 8
Sample line ! 1 Seconds
Oven TeMp : 168 °C
Chart Length ! 15 Minutes
Detector :E,C,D,
ColiiMn !CAP18'i
Pressure : 28 psi
PEAK RESULT BHEC
PCB 1 58,08 PPM ?8
PCB 2 50,00 PPH 13?
PCB 3 58,80 PPH 161
PCB 4 58,08 PPH 239
11
-------
** SCIX'OCJiAPK ** H
83-25-1988 13:52:02 60AROC1242 100 I
43-29-1988 13:16:54 69 7 '/,
|w,yw/_* ViV
Figure 7
14
-------
IDENTIFICATION OF SEMI VOLATILE ORGANIC COMPOUNDS IN SELECTED AIR
SAMPLE EXTRACTS BY GAS CHROMATOGRAPHY/MATRIX ISOLATION INFRARED
SPECTROMETRY
JeffreyW.Childers
Northrop Services, Inc. - Environmental Sciences
P O. Box 12313
Research Triangle Park, North Carolina 27709
Nancy K. Wilson and Ruth K. Barbour
[J- S. EPA, Environmental Monitoring Systems Laboratory
Research Triangle Park, North Carolina 27711
• The.use °' 9as chromatography/matrix isolation infrared (GC/MI-IR) spectrometry for the
identification of semivolatile organic compounds in environmental air sample extracts is
I monstrated. Several polycyclic aromatic hydrocarbons (PAHs), including oxygenated PAH and
aikylated PAH compounds, were identified in an extract of paniculate emissions from a wood-
ournmg stove. The potential of GC/MI-IR spectrometry for quantitative analyses is discussed.
15
-------
IDENTIFICATION OF SEMIVOLATILE ORGANIC COMPOUNDS IN SELECTED AIR SAMPLE EXTRACTS
BY GAS CHROMATOGRAPHY/MATRIX ISOLATION INFRARED SPECTROMETRY
Introduction
Current efforts in our laboratory involve the development and evaluation of gas
chromatography/matrix isolation infrared (GC/MI-IR) spectrometry for the characterization of
semivolatile organic compounds in air. In, GC/MI-IR spectrometry, the components that are
eluted from the GC are trapped in an inert matrix as it is deposited on the surface of a rotating,
gold-plated cryogenic disk.1 The analytes remain frozen on the disk indefinitely and can be
examined by MI-IR spectrometry after completion of the GC run. This allows IR scans to be co-
added as necessary to increase the signal-to-noise ratio to an acceptable level. Also, because
each trapped molecule is surrounded only by inert matrix atoms (e.g., argon) at 14 K,
intermolecular interactions and molecular rotations that cause IR band broadening are
eliminated. As a result, MI-IR spectra exhibit sharp, well-defined features. This enhances the
ability of GC/MI-IR spectrometry to distinguish between isomers and other closely related
compounds.
One important class of compounds that constitutes a significant fraction of the
semivolatile component of environmental air samples is the polycyclic aromatic hydrocarbons
(PAHs). We have previously shown the ability of GC/MI-IR spectrometry to discriminate between
isomeric PAHs such as chrysene and triphenylene, the benzofluoranthene isomers, and benzole]-
and benzo[a]pyrene in extracts from urban air paniculate matter and diesel-powered
automobile emissions.2 The ability to discriminate between isomers is crucial in assessing the
health risk associated with air sample extracts because of the vastly different biological
activities, such as carcinogenicity and mutagenicity, often exhibited by isomeric compounds. In
this paper, preliminary results from the use of GC/MI-IR spectrometry for identifying PAH
compounds in an extract of wood-burning stove particulate emissions are presented.
Experimental Methods
All results were obtained on a Mattson Instruments (Madison, Wl) Cryolect system. The
Cryolect system consists of a Mattson Instruments Sirius 100 Fourier transform IR spectrometer
and a Hewlett-Packard (Palo Alto, CA) 5890A gas chromatograph, equipped with a flame
ionization detector, interfaced to a Mattson Instruments cryogenic module. The Cryolect system
is described in detail by Reedyetal.3
MI-IR spectra were derived from 128 co-added, double-sided, 8192-point interferograms.
A zero-filling factor of 2 and a triangular apodization function were applied to the co-added
interferograms before the fast Fourier transform was performed. This resulted in MI-IR spectra
with a nominal resolution of 4 cnV. Single-beam sample spectra were collected with the
cryogenic disk positioned at the GC peak maximum, and single-beam background spectra were
collected with the disk positioned 0.30 min before the GC peak maximum. The ratios of the
single-beam sample spectra to the single-beam background spectra were plotted as absorbance
files.
The analytical column was a 30-m x 0.25-mm i.d. DB-5 fused-silica capillary with a 0.25-um
film thickness (J&W Scientific, Folsom, CA). The effluent of the column was split; 20% of the
effluent was directed to the FID, and the remaining 80% was directed through an open-split
cross, and then through a heated fused-silica transfer line to the cryogenic disk. The carrier gas
was 0.62% argon in helium (National Welders, Raleigh, NC). The GC oven was held at an initial
temperature of 40 °C for 4 min, increased at 20 °C/min to 200 °C and held for 1 min, then
increased at a rate of 2 °CAnin to 300 °C and held for 10 min. The FID heating block, the transfer
line conduit, and the deposition tip heating block were held at 300 °C. The cryogenic disk was
maintained at 14 K throughout matrix deposition and MI-IR spectral acquisition. The pressure in
the cryogenic chamber was 1.5x 10"6torr.
16
-------
Results and Discussion
In developing the methodology for the analysis of environmental air samples by GC/MI-IR
spectrometry, selected sample extracts generated under the Integrated Air Cancer Project (IACP)
have been analyzed. The IACP is a long-term, interdisciplinary program whose goals are to
identify the principal carcinogens in air, determine the major sources of these carcinogens, and
improve the estimate of human exposure and health risk from specific emission sources.4
Selected samples collected as part of an IACP pilot study conducted in Albuquerque, NM, from
January to March 19855 were analyzed by GOMI-IR spectrometry.
As part of the Albuquerque pilot study, paniculate matter was collected from the
combustion of Pinyon pine in a wood-burning stove. Sample fractions from the extract of the
stove emissions were generated through bioassay-directed fractionation.6 Particulate matter
collected from the stove emissions was extracted with methylene chloride, then fractionated by
solvent partitioning in 10% methylene chloride/hexane, methylene chloride, and methanol. The
methylene chloride partition fraction was separated into three fractions {nonpolar, moderately
polar, and polar) by step-gradient high-performance liquid chromatography (HPLC).7
A portion of the GOFID trace for the nonpolar HPLC fraction of the methylene chloride
?ction of thie solvent-partitioned extract is shown in Figure 1. Several PAH compounds were
identified by GC/MI-IR spectrometry in this fraction. Figure 2A shows the MI-IR spectrum of the
component that eluted at 18.83 min. This spectrum illustrates the sharp spectral features and
the excellent signal-to-noise ratio typically exhibited by MI-IR spectra, even for relatively minor
components in the chromatogram. Although a quantitative analysis was not performed on this
fraction, the concentration of this component is estimated to be approximately 2 ng/uJ on-
column. A search for this spectrum in the reference MI-IR spectral library identified this
component as anthraquinone (compare Figures 2A and 2B). Likewise, the component that
fluted at 24.26 min was identified as l-methyl-7-isopropylprienanthrene, retene (compare
figures 3A and 38). Retene is a major component of wood-burning stove emissions and has
been proposed as a marker compound for source apportionment.8 The following PAHs were
also identified in this fraction: 9-fluorenone (15.35 min), phenanthrene (15.93 min),
fluoranthene (20.69 min), pyrene (21.84 min), benz[a]anthracene (30,70 min), chrysene and
tnphenylene (31.03 min), benzo[e]pyrene (42.89 min), and benzo[a]pyrene (43.29 min).
Oiethylphthalate (13.58 min) and bis(2-ethylhexyl)phthalate (33.85 min) were also identified in
this fraction. The identified components represent more than 25% of the total integrated GC
Peak area. Contrary to the expectation that most PAHs would be extracted by the nonpolar-
solvent-partioning step, several PAHs were found in the nonpolar HPLC fraction of the
methylene chloride partition. In fact, when compared to previous results.2 this fraction was
found to contain many of the same PAHs as were in the nonpolar-solvent-partitioned extract.
Future studies will include the comparison of paniculate samples collected from
w°oa$rnoke-im parted airsheds with the emissions from wood-burning stoves. However, for this
jype of comparison to be meaningful, the quantitative aspects of GC/IVH-IR spectrometry must
T|rst be addressed. In GOMI-IR spectrometry, there are two potential sources for quantitative
results: the FID response and the MI-IR spectral response. Preliminary studies indicate that, with
the use of an internal standard, the FID response for target PAHs is very reproducible (<1.5%
relative standard deviation). For the MI-IR spectral response, the ratio of the MI-IR peak
maximum intensity to the GC/FID peak area should also be constant, provided that the portion
the GC effluent that is split between the FID and cryogenic disk is held constant. However, in
replicate analyses of a standard solution containing target PAHs, this ratio showed a relative
standard deviation greater than 15%. Efforts are currently under way to improve the
reproducibility of the MI-JR spectral results.
inclusions
.In preliminary studies, we have shown that GGMI-IR spectrometry combines the
erisitivity needed to detect compounds at the low levels typically found in environmental
»mples with the specificity to discriminate between isomers and other closely related
Orr»pouncts. To date, GC/MI-IR spectrometry has provided the qualitative identification of
17
-------
target PAHs in several air sample extracts. The development of GC/MI-IR spectrometry to include
quantitative analyses will provide for the unambiguous determination of compounds in air that
are suspected to pose a significant risk to human health.
References
1. S. Bourne, G. Reedy, P. Coffey, D. Mattson, "Matrix isolation GOFTIR," Am. Lab.. J6: 90 (1984).
2. J.W. Childers, N.K. Wilson, and R.K. Barbour, "Application of matrix isolation infrared
spectrometry to analysis for polynuclear aromatic hydrocarbons in environmental samples," in
Proceedings of the Eleventh Annual International Symposium on Polynuciear Aromatk
Hydrocarbons. Gaithersburg, MD, Sept. 23-25, 1987.
3. G.T. Reedy, 5. Bourne, P.T. Cunningham, "Gas chromatography/infrared matrix isolation
spectrometry," Anal.Chem.. 51.: 1535 (1979).
4. J. Lewtas, L. Cupitt, "Overview of the Integrated Air Cancer Project," in Proceedings of the
1987 EPA/APCA Symposium on Measurement of Toxic and Related Air Pollutants, Air Pollution
Control Association, Pittsburgh, PA (1987), p. 555.
5. V.R. Highsmith, C.E. Rodes, R.B. Zweidinger, R.G. Merrill, "The collection of neighborhood air
samples impacted by residential wood combustion in Raleigh, NC and Albuquerque, NM," in
Proceedings of the 1987 EPA/APCA Symposium on Measurement of Toxic and Related Air
Pollutants, Air Pollution Control Association, Pittsburgh, PA (1987), p. 562.
6. D. Schuetzle, J. Lewtas, "Bioassay-directed chemical analysis in environmental research," Anal.
Chem..58: 1060(1986).
7. Ray Merrill, Radian, Research Triangle Park, NC, personal communication (1987).
8. T. Ramdahl, "Retene: A molecular marker of wood combustion in ambient air," Nature. 306:
580(1983).
18
-------
CO
Minutes
Figure 1 A portion of the GOFID trace for the tACP nonpolar HPLC fraction of the methylene chloride partition extract. Components
identified include (1) diethylphthalate, (2)9-fluorenone, (3) phenanthrene, (4) anthraquinone, (5) fluoranthene, (6) pyrene.
(7)retene, (8) benz[a]anthracene, (9) chrysene/triphenylene, (10) bis(2-ethythexyt)-phthalate, (11)benzole]pyrene, and
(12) benzo[a]pyrene.
-------
Mlcror*
5.0
5.5
I
8.0
I
6.5
I
7.0
B.O
I
9.0
I
10.0
14
1
B
2000
1BOO
I
I
1400
1200
1000
BOO
Ifavenumber
Figure 2 MI-IR spectrum of (A) the component of the IACP nonpolar HPLC fraction with
retention time of 18.83 min and (B) anthraquinone reference standard.
Microns
IBM
1400
Havenumber
Figure 3 MI-IR spectrum of (A) the component of the IACP nonpolar HPLC fraction with
retention time of 24.26 min and (B) retene reference standard.
20
-------
COUPLED SUPERCRITICAL FLUID EXTRACTION/GAS
CHROMATOGRAPHIC ANALYSIS OF TRACE ORGANICS
FROM ATMOSPHERIC SAMPLES
Steven B. Hawthorne, Mark S. Krieger, and David J. Miller
University of North Dakota
Energy and Mineral Research Center
Grand Forks, North Dakota, 58202
Supercritical fluid extraction (SFE) can yield rapid and quantitative
recovery of trace organic compounds from atmospheric particulates and from
sorbent resins. Since many supercritical fluids (e.g., (X^) are gases at
room temperature and pressure, the SFE step can also be directly coupled with
capillary gas chromatography by simply inserting the outlet restrictor of the
*E cell into the gas chromatographic column through a standard on-column
injector (coupled SFE-GC). SFE-GC yields maximum sensitivity since the
extracted species are quantitatively transferred into the GC column where
they are cryogenically trapped prior to normal gas chromatographic analysis
using MS, FID, or ECD detection. With the use of the coupled SFE-GC
•echnique, all of the analysis steps including analyte extraction,
concentration, and gas chromatography can be completed in a total time of
less than one hour. The use of SFE-GC analysis for the rapid extraction,
ientification, and quantitation of trace atmospheric organics from air-borne
Particulates and from Tenax sorbent resin is presented.
21
-------
Introduction
The extraction of toxic and related air pollutants that are associated
with air-borne participates or that have been collected on sorbent resins is
typically performed using lengthy liquid solvent extraction techniques.
Extraction with liquid solvents often requires several hours or even days to
perform, does not always result in quantitative recovery of analyte
compounds, and results in a sample that is diluted in a large volume of
solvent. Even with concentration of the extract to a small (e.g., 100 ;iL)
volume, the majority of the collected analytes are generally discarded since
gas chromatographic injections can only acconmodate approximately 1 jiL. The
use of thermal desorption of sorbent resins is also useful for analytes
collected on sorbents such as Tenax, but this technique can lead to thermal
degradation and/or incomplete recoveries of analyte species.
Supercritical fluid extractions (SFE) have been shown to yield rapid and
quantitative extraction and recovery of a variety of organic pollutants from
samples ranging from air-borne particulates to river sediments. We have
developed a method for directly coupling the supercritical fluid extraction
step with capillary gas chromatography (SFE-GC) which requires no
modification of commercially available GC instrumentation*. Since every
analyte molecule is transferred directly into the GC capillary column,
maximum sensitivity is achieved which dramatically reduces the quantity of
sample that must be collected. Coupled SFE-GC has the additional advantage
of obtaining quantitative extraction and recovery of organic pollutants from
a variety of sample matrices in approximately ten minutes, and the
extraction, sample concentration, and GC analysis can be completed in a total
time of less than one hour.
Experimental Methods
All GC/MS analyses were performed using a Hewlett-Packard model 5985
GC/MS. GC/FID analyses were performed using a Hewlett-Packard model 5730 or
model 5890 gas chromatograph. All of the gas chromatographs were equipped
with a standard on-column injector of the type supplied with the model 5890,
and carrier gas flows were identical to those supplied with the model 5890
(i.e., back-pressure regulated). All capillary gas chromatographic columns
were supplied by J and W Scientific and had DB-5 as the stationary phase.
Both wide-bore (30m X 0.32 mm i.d., 1 ^jm film thickness) and narrow-bore (60m
X 0.25 mm i.d., 0.25 /im film-thickness) were used.
The method for coupling the SFE step with the GC is shown in Figure 1.
The extraction cell was constructed using "Parker" brand 316 stainless steel
fittings (as described in reference 1) and had an internal volume of
approximately 0.1 mL. Supercritical pressures were maintained inside the
extraction cell by using 15-cm lengths of capillary fused silica tubing
(Polymicro Technologies, Phoenix, AZ, USA) for outlet restrictors. Flow rate
through the extraction cell was controlled by using tubing (150 pn o.d.) with
internal diameters of ca. 20 /m (for narrow-bore GC columns) or ca. 25 jm
(for wide-bore GC columns). Temperature was maintained at 45 °C by inserting
the extraction cell into a thermostated tube heater.
The direct coupling of the SFE cell to the GC column was accomplished by
inserting the SFE outlet restrictor capillary into the gas chromatographic
column through the on-column injection port. The gas chromatographic oven
was cooled during the extraction to cause the extracted analytes to be
thermally focused inside the chromatographic column at the outlet of the SFE
22
-------
restrictor. SFE-GC analyses were performed using the following steps:
Firstj the extraction cell was placed inside of the tube heater (held at 45
°C) and the outlet restrictor was inserted into the GC column through the on-
column injection port. The cell was then pressurized with the supercritical
C02 and the extraction was allowed to proceed for ten minutes. The GC oven
temperature was held constant (-30 to 5 °C) during the entire extraction
step. After the extraction was completed the SFE restrictor was withdrawn
from the on-column injector, the C02 was allowed to flush from the GC column
for two minutes, and the gas chromatographic analysis was performed in a
normal manner.
Results
The ability of the coupled SFE-GC technique to yield good chromatographic
peak shapes and rapid and quantitative extraction of urban air organics (4
liter air sample) from an 80 mg Tenax-TA trap is shown in Figure 2. The
Tenax was extracted for 15 minutes with 200 atm C(>2. The lower chromatogram
shows the second SFE-GC analysis of the same Tenax trap. The lack of
significant species in the second extract indicates that the first 10 minute
extraction was sufficient to quantitatively recover the trapped analytes.
Quantitative SFE extraction of PAHs ranging from naphthalene (mol. wt.=128)
to coronene (mol. wt.*300) from Tenax-GC in an extraction time of 15 minutes
has also been demonstrated as shown in Table I. It is important to note that
the temperature during all of these extractions was only 45 °C.
Coupled SFE-GC analysis using supercritical ^0 extraction for 15 minutes
has also been applied to the extraction and GC/MS quantitation of PAHs from
urban dust (National Bureau of Standards SRM 1649). The SFE-GC/MS analysis
required only two mg of sample, and less than one hour to perform per sample
including the extraction, concentration, and GC/MS analysis steps. In
contrast, the NBS method required one gram of sample and 48 hours of Soxhlet
extraction followed by several concentration and class-fractionation steps.
As shown in Table II, the results of SFE-GC/MS were in excellent agreement
with the certified values.
Coupled SFE-GC/FID analysis of wood smoke particulates is shown in Figure
3. As shown in the lower chromatogram, a 10 minute extraction with 300 atm
C02 was not sufficient to quantitatively extract the more polar phenolic
species. Longer extraction times and/or the use of supercritical fluids that
are more polar than C02 may be required to obtain quantitative recovery of
organics which have higher polarities. The SFE-GC/MS analysis of volatile
organics from a spruce needle is shown in Figure 4. Even the volatile
•nonoterpenes show good chromatographic peak shape, and the high concentration
of water (ca. 80% by weight) did not appear to adversely affect the
extraction and chromatography.
Conclusions
Supercritical fluid extractions are a powerful alternative to liquid
solvent extractions for the recovery of organic pollutants from particulates
and sorbent resins. Coupled SFE-GC analysis can be used to perform
quantitative extraction, sample concentration, and GC analysis in a total
time of less than one hour and, since all of extracted analytes are
quantitatively transferred in to the GC column, the amount of sample that
needs to be collected is dramatically reduced.
23
-------
Acknowledgements
The financial support of the U.S. Environmental Protection Agency, Office
of Exploratory Research (grant no. R-812229-01-1) is gratefully acknowledged.
References
1. S.B. Hawthorne and D.J. Miller, "Directly Coupled Supercritical Fluid
Extraction-Gas Chromatographic Analysis of Polycyclic Aromatic
Hydrocarbons and Polychlorinated Biphenyls from Environmental Solids," J_._
Chromatogr. 403 63-76 (1987).
Table I
Supercritical 002 Extraction of PAHs from lenax-GC
Recovery
Species mol. wt. Trial 1 Trial 2
naphthalene 1ZB 101 100
9-fluorenone 180 96 95
phenanthrene 178 97 97
pyrene 202 98 98
benzta]anthracene 228 96 98
benzo[ghi]perylene 276 94 110
coronene 300 103 93
Table II
Coupled SFE-GC/MS Analysis of FAHs from NBS
Standard Reference Material 1649 (Urban Dust)
Concentration (ug/g)
Certified Coupled
Valuea SFEH3C/MSb
fluoranthene 7.1+0.5 7.3±1.0
benz( a] anthracene 2.6+0.3 2.6±0.8
benzol a] pyrene 2.9± 0.5 2.8±0.5
benzo[ghi]perylene 4.5± 1.1 3.6+0.9
indeno[1,2,3-cd]pyrene 3.3± 0.5 3.0± 0.5
aValue certified by the National Bureau of Standards.
bBaaed on four replicate analyses of 2-cqg samples.
24
-------
Pump
Extraction Cell
Tube Heater 45 °C
Outlet Restricted
On-Column Injection Port
Oven Wall
-30 to 5 °C
Capillary
GO Column
Detector
Figure 1 Diagram of method used for coupled SFE-GC analysis.
individual analysis steps used are described in the text.
The
CD
CO
o
a
w
CD
DC
Coupled SFE/GC of Urban Air Organics
Trapped on Tenax-TA
20
40
Retention Time (min)
Figure 2 Coupled SFE-GC analysis of a 4 L urban air sample collected on
Tenax-lA. The top and bottom chromatograms show the first and second SFE-GC
analyses, respectively.
25
-------
Wood Smoke Particulates
1st extraction
0
OT
C
o
Q.
W
(D
DC
LL
Retention Time (min)
Figure 3 Coupled SFE-GC analysis of wood smoke particulates. The top and
bottom chromatograms show the first and second SFE-GC analyses, respectively.
Spruce Needle
c
CD
O
c
o
(2
/3-pinene
camphene
a-pinene
Mimonene
'bornyl acetate
j»'wj\>^^.., ..J^N
5 10 15
Retention Time (min)
Figure 4 SFE-GC/MS analysis of volatile organics from a spruce needle.
Extraction conditions are given in the text.
26
-------
PHOTOCHEMICAL AGING OF POLYCYCLIC AROMATIC
HYDROCARBONS FOUND IN JET ENGINE EXHAUST
by
Michael R. Kuhlman, Jane C. Chuang, and Surendra B. Joshi*
BATTELLE
Columbus Division
505 King Avenue
Columbus, Ohio 43201-2693
I. Introduction
Engineering and Services Center of the United States Air Force
is conducting a joint program with the Naval Air Propulsion
-------
it was selected as providing the maximum information possible within the
available resources.
Sampling of the engine exhaust was performed for two purposes:
injection of a measured volume of engine exhaust into the chambers for
the aging experiments, and collection of exhaust components on filters
and sorbent media for subsequent analysis. All sampling was performed
by withdrawing exhaust from a point located near the exhaust plane of
the engine, near the longitudinal axis of the engine. To permit engine
exhaust characterization to be performed at the same time as the chamber
filling was performed, an auxiliary 3/8" I.D. stainless steel sampling
probe was affixed to the cruciform sampling rake. The inlet for this
probe was located in the same plane as the inlets for the rake. This
probe was connected to a clean air purge line to prevent contamination
from entering the probe until sampling was initiated. Also connected to
the probe was a 50' length of 3/8" I.D., heated traced Teflon tubing
which was maintained at 50 C. This Teflon line conveyed the exhaust to
a heated stainless steel metal bellows pump, which provided a flow rate
of 3 scfm. This pump was used during both the exhaust sampling operation
and during the chamber filling operation.
The three smog chambers were constructed as right circular cylinders
having a diameter of 3.0 m and a height of 3.4 m, with a conical top of
height 0.6 m affixed to their top. The total volume enclosed by the
chambers was 26.7 nH. The materials used to construct the chambers con-
sisted of an aluminum framework, stainless steel hardware, and Teflon
film. One chamber was covered with an aluminized mylar film to exclude
sunlight for those experiments which were to be performed in darkness.
The chambers were equipped with an axially positioned mixing fan,
sampling/injection ports which entered through the chamber floor, and an
exhaust blower. The chambers were fastened to wooden platforms, approxi-
mately 3.7 m square, equipped with wheels and a steering yoke which
enabled them to be pulled by hand or towed with a vehicle at low speeds.
Photochemistry Measurements
During the photochemical aging experiments, the concentrations of
total hydrocarbons (THC), 03, NO, NOX, and SF§ within each of the cham-
bers was monitored. Heated Teflon sampling lines were used to connect
the three chambers to a switching station which directed air from one of
the chambers or ambient air into the sampling manifold to which the
instrumentation was connected. The data acquisition program cycled the
sampling through each of the four sources such that each was monitored
for 5 minutes in sequence.
Overnight prior to performance of an aging experiment the chambers
were purged either with house air which was filtered and run through an
activated charcoal scrubbing apparatus. Upon arrival onsite the purge
air was turned off and SF§ injections were made into each chamber while
the background concentrations of all parameters were obtained. After
the concentration of SF5 was established, chambers 1 and 3 were reposi-
tioned in preparation for engine exhaust injection. Following exhaust
injection, each chamber was returned to the site for the aging experiment
and immediately connected to the sampling system.
At several times during the aging experiments samples of the cham-
bers were collected on filter and sorbent media for subsequent analysis
28
-------
for the target compounds analyzed using 6C/MS techniques. At the conclu-
sion of the aging experiment a large volume sample (-5 m3) of the chamber
atmosphere was collected in order to assure collection of as much mass
of the PAH and N02-PAH species as possible.
HI. Results
The target PAH compounds which are biologically active (benzo[a]>
anthracene, cyclopenta[c,d]pyrene, benzofluoranthenes, benzo[a]pyrene,
indeno[l,2,3,-c,d]pyrene, and dibenzo[a,h]anthracene) are only present
near or below detection limits in any of the samples obtained. In no
instance is there any increase in concentration of these compounds during
an aging experiment, as one would expect.
As a rule, the target PAH all appear to decay during the aging
experiments, at a greater rate in sunlight than in the dark. The behav-
ior of naphthalene, the most prevalent of the target PAH compounds in
tnese measurements, seen in Figure 1, is typical of the class. Phenan-
tnrene concentration data are shown for selected runs in Figure 2. This
compound is the only one of the PAH which exhibited an increase in
concentration at any point in the runs performed. These small apparent
increases in concentration are almost certainly due to imprecision in
tne sampling and analysis of these compounds.
. Wn11e it will be seen below that the very active system, TF33-P7
th IS at 30* power with added EKMA m1x» led to formation of several of
Sin f comP°unds» the effect on the PAH compounds is not immediately
evident, as both chambers exhibit similar decay profiles for the measured
compounds. This is not surprising since the concentrations of NOg-PAH
?nnrrv! *re ver* low " these are expressed in ng/m3, whereas the PAH
Concentrations are expressed in ng/m3. Because of this disparity in
oncentrations, one cannot observe any change in the PAH parent compound
concentration which corresponds to the formation of the derived NOg-PAH.
st T!je distribution of PAH between vapor and particulate matter is
nanhJS i dotnir»ated by the vapor phase. The most abundant compound,
ipE ,, ne' 1s found on1y 1n the vaP°r Phase, as is true for acenaphtha-
While the pyrene is initially found associated with the filter
, it is present mainly in the vapor phase at the end of the runs.
S "°J; known how much of this shift is due to loss of aerosol from
ma Fnamber atmosphere and how much is due to chemical processes which
may be occurring as the test atmosphere ages.
Examining the biologically active NOg-PAH compounds first: 3-nitro-
(3-NFJ, 1-nitropyrene (1-NP), and dinitropyrene, we note
Who I!uS last compound was detected in none of the samples collected.
fni *• e is almost no 3-NF found in the J79 runs and certainly no
formation, the TF33 engines 30 and 75 percent power exhaust results
' formation of this compound during the middle of the test run and its
iuosequent decay. Figure 3 depicts 3-NF concentrations in the TF33
in ?"?*• When the EKMA mix is added to the engine exhaust, it results
fU *!nced 3"NF formation. It is interesting that the idle exhaust
or • IS TF33 en9^ne does not result in formation of 3-NF, in sunlight
fLJ" th5 dark' Similar results are found for 1-NP, in that formation
isnnanced by the EKMA mix) and subsequent decay are observed, to differ-
ent degrees for different runs.
29
-------
The compound 2-nitro-l-naphthol is present at the highest concentra
tion of the measured N02-PAH in all cases. It is also seen to decay in
nearly every run, doing so at a greater rate in sunlight than in the
dark, suggesting possible photolysis. For the run using TF33-P3, 75
percent power exhaust this compound does not appear to decay, but shows
a slight increase throughout the run. This chamber contained the lowest
initial concentration of this compound. The 3Q% power exhaust, with
added EKMA resulted in a doubling of this compound's concentration,
followed by a strong decay after noon. In every case having sufficient
data, 1-nitronaphthelene is formed during the mid-day period and then
decays. Figure 4 illustrates this compound's behavior under several
sets of conditions.
To summarize the NOg-PAH data, several comments can be made. In
most instances, the relative concentrations of the major NOg-PAH
compounds measured do not vary by more than approximately a factor of
five in the initial conditions for the chamber. The fact that some of
these compounds are observed to form in the chambers while others do not
changes their relative abundance. None of these compounds measured is
stable through the course of the experiments, and the presence of sun-
light increases the rate of decay of several of the compounds. In a
number of cases, certain compounds' final and initial concentrations in
the chamber were nearly the same, but when this occurred, 1t was due to
formation and decay of the compound during the run.
IV. Conclusions
The effects of engine type, power setting, sunlight, and photochemi
cal activity of the atmosphere have been investigated in this study
through use of matched reaction chambers which received injections of
jet engine exhaust. Several conclusions can be drawn from the measure-
ments made of the engine exhaust samples and from the measurements
performed on the chambers as the exhaust containing atmosphere within
them aged.
Conclusions which can be offered based upon the measurements of the
engine exhaust are the following:
• The inorganic components of the exhaust particulate emissions
are negligible under all conditions examined.
t The THC content of the exhaust decreases as engine operating
power increases, such that at 75 percent power, the exhaust THC
approximates the background air concentration.
• The NOX concentration and NO:NOX ratio Increase with increasing
power 1n a similar fashion for all three engines.
t Most PAH in the engine exhaust track the concentration of THC.
As exceptions, the fluoranthene and pyrene persist 1n the TF33
exhaust at 30 percent power.
• The NC-2-PAH exhaust concentrations track the emitted THC less
well, due to the role of N02 in their formation. Many of these
compounds are present 1n the 30 percent exhaust at their greatest
concentration.
30
-------
• The exhaust from the engine operated at 75 percent power contains
so little hydrocarbon it is probably not worth examining in
future studies of this sort.
Regarding operational aspects of this study, the following conclu-
sions are offered:
• The portable smog chambers designed and constructed for this
study served their function well in all respects.
• The low concentration of many of the compounds of interest in
the exhaust requires that large sample volumes be handled. Any
future studies of this sort should include sampling equipment
capable of drawing larger flow rates (e.g., 1 m3/min). This is
especially true when sampling the exhaust at higher power set-
tings.
The photochemical aging of the exhaust provided data from which the
TOIlowing conclusions can be drawn:
• Biologically active 5- to 6-ring PAH are found at low levels in
the exhaust from the TF33 engines at idle. Under all other
conditions these compounds are at or below detection limits in
the samples collected.
• Biologically active N02-PAH compounds are found at low levels in
each of the exhaust samples collected. Formation of two of these
compounds was observed during the aging experiments.
• The PAH and NO?-PAH measured are principally found in the vapor
state.
• PAH are observed to decay under all conditions examined, the
rate of decay being greater in sunlight than in the dark. No
consistent PAH formation was observed.
• Several NOg-PAH compounds were observed to form during the mid-
day portion of the tests, especially in photochemically reactive
atmospheres.
• For initial conditions of high THC:NOX ratio, such as Idle
exhaust Injected into chambers, very rapid conversion of NO
appears to have taken place 1n the chambers, without subsequent
03 accumulation.
31
-------
woo
•00
I 400
O
100
CO
ro
KM)
Figure 1
o.
O--..
o
TF33-P7;IDL£;L
Tf33-P7;30%+EKMA;L
..-O
O
J79C:JPLE;L __
TF33-P7;30S;L
TF33-P7;30%:L
10
Q 14
TIME (CDT)
18
10
20
Naphthalene concentration profiles
for selected runs (engine, power
level, and light or dark aging
as indicated^.
o>
3
D
C
O
t>
en
33
*
c
ff
2-
a
c
J79C:JDLE;0
TF33-P7;IDLE;L
TF33-P7;30R-i-EKMA;L
O,
J79C:JOLE:L
TF33-P7;30%;L
TF33-P7;30%;L
.0.
10
12 14
TIME (CDT)
18
20
Figure 2.
Phenanthrene concentration profiles
for selected runs (engine, power
level, and light or dark aging
as indicated).
-------
US
CO
TF33-P3;75K;L
TF33-P7;30X;L
12 14
TIME (CDT)
Figure 3. Concentration profile of 3-nitrofluor-
anthene for selected runs (engine, power
level, and light or dark aging as
indicated)*
TF33rP3jlpLE;p
TF3J-P7;30Z+EKMA;L
TF33-P3;75!S;L
TF33-P7;30?S;L
12 U
TIME (CDT)
Figure 4. Concentration profile of 1-nitro-
naphthalene for selected runs (engine,
power level, and light or dark aging
as indicated).
-------
ANALYTICAL AND SAMPLING METHODS OF THE NONOCCUPATIONAL
PESTICIDE EXPOSURE STUDY (NOPES)
J. P. Hsu, H. G. Wheeler, H. J. Schattenberg III,
P. V. Kuhrt, H. J. Harding and D. E. Camann
Southwest Research Institute
P.O. Drawer 28510
San Antonio, Texas 78284
An analytical protocol was developed to analyze for 33 pesticides in
indoor, outdoor, and personal air samples, drinking water, and on gloves
worn to assess dermal exposure. Soxhlet extraction was used for the
polyurethane foam plugs employed for air sampling and for the gloves. The
extraction procedure of U.S. EPA Method 608 was used for the water
samples. A gas chromatography/electron capture detection (GC/ECD) and gas
ehromatography/mass spectroscopy/multiple ion detection (GC/MS/MID)
analytical approach was used for analysis.
34
-------
Introduction
Many Insecticides, fungicides, and herbicides are used in and around
houses to control pests. Human exposure to the pesticides were measured
through air sampling, dermal contact and drinking water. This study
included sampling and analysis for 33 pesticides (Table I) in approx 260
different homes in Jacksonville, FL, and Springfield, MA.
Experimental
Cleanup of PUFs and Gloves
The polyurethane foam (PUF) plugs (22-mm OD, 7.6-cm length) or cotton
gloves were cleaned by using Soxhlet extraction with acetone and then 6Z
diethyl ether in n-hexane. Each extraction was performed for 48 h at 50
min/cycle. After extraction, the PUF plugs or gloves were placed in a
precleaned, capped cylindrical vessel, and the residual solvent was
removed by passing charcoal-treated nitrogen through the vessel. One PUF
plug or glove from each clean-up batch was extracted and analyzed to make
sure no targeted pesticides were found above their quantitation limit
goals. Each PUF plug was then installed into a precleaned glass sampling
cartridge. The PUF cartridges and glove pairs were wrapped with hexane-
rinsed aluminum foil and stored in precleaned jars for shipping.
Sampling
The air sampling procedure developed by Lewis, et al.1 was utilized
to collect airborne pesticides in each household at indoor and outdoor
locations of daily activity and for personal air monitoring. The samples
were collected by drawing ambient air through FUF cartridges at a flow
rate of approximately 3.8 1/min for 24 h with a Du Pont Alpha 1 pump. The
flow rate was checked at the beginning and the end of the sampling using
an electronic bubble tube calibrator. A black plastic sleeve covered the
cartridge during sampling and a short piece of Tygon tubing connected the
cartridge to the pump inlet. After sampling, the PUF cartridge was
wrapped with aluminum foil and returned to the original jar for shipment
to the laboratory. Dermal exposure was determined by having participants
wear a pair of clean gloves during a pesticide application. A concurrent
personal air sample was also collected. Both the air and dermal samples
were kept cool by dry ice until extraction. Water samples were collected
from the household drinking water source, typically the kitchen faucet.
The water samples were maintained at 4'C until extraction.
Extraction
The PUF plug or glove sample, after spiking with 100 pi of a solution
°f 1.2 ng/jil of octachloronaphthalene as a recovery surrogate, was
Soxhlet-extracted with 300 ml of 6Z diethyl ether in hexane for a minimum
°f 16 hours. The extract was poured through a tube packed with anhydrous
sodium sulfate and concentrated in Kuderna-Danish apparatus to 10 ml, The
extract was split in half and further concentrated to 1 and 0.5 ml by
nitrogen blovdown for 6C/ECD and GC/MS/MID analysis, respectively. Vith
e*ch extraction batch of 20-30 samples, two clean PUF plugs were spiked
with a 100 fil of a standard (matrix spike) solution at the concentrations
given in Table II. The spiked PUFs were treated as samples. A clean
glove was spiked with a 100 pi of the matrix spike solution for every
seven glove samples received. For each extraction batch of PUF plugs or
gloves, a solvent blank was also prepared by spiking 100 ,ul of surrogate
solution into a Soxhlet extractor containing SO ml of extraction solvent.
35
-------
The rest of the blank extraction was the same as that for samples. The
water samples were extracted according to EPA Method 60S;2 the final
extract split and volumes were identical to those of PUFs. For every
seven water samples, a water sample was spiked with 100 nl of matrix spike
solution and extracted along with the samples. All the extractions were
completed within seven days after sample collection.
6C/ECD Analysis
A HP5890 GC with dual injectors and ECDs, equipped with a HP7673
autosampler was used. Both a 15-m X 0.53-mm ID OV-17 bonded fused silica
column and a 30-m X 0.53-mm ID DB-5 bonded fused silica column were
installed in the GC for analyses. The GC was set at 60"c for two minutes
then programmed to 1AO°C at 25°C/min, then programmed to 270°C at 4°C/min>
The temperature of both injectors was 235°C and detectors 350°C. The
helium flow rate was 10 ml/min and the volume of each injection 2 /il. A
dynamic range was generated at the beginning of each sampling season fot
both analytical columns. The dynamic range covered 5 to 80 or 100 times
the quantitation limits. In most cases, all GC/ECD targeted compounds,
except pentachlorophenol and dicofol, showed less than 252 relative
standard deviation (RSD) over the concentration range. These two
compounds usually showed inconsistently large RSDs.
An analytical sequence started with the injections of three linearity
check solutions A, B and C (Table III). The linearity of the GC/ECD wad
checked by calculating the RSD of the three response factors of each
compound in Table III. The RSDs for both analytical columns were belotf
102 in most cases. The GC/ECD single-component targeted compounds*
divided into four different individual standard mixtures I, II, III and IV
(Table I), were injected after the linearity check. The entire analytical
sequence was:
linearity check solutions A, B, and C,
individual standard mixtures I, II, III, and IV,
chlordane standard (8 ng per injection),
solvent blank,
half of samples in extraction batch,
individual standard mixtures I, II, III and IV,
other half of samples in extraction batch,
individual standard mixtures I, II, III and IV
The retention time window on both columns of every single-component
compound was determined for each analytical sequence. Three times the
standard deviation of the retention times from the three different
standard runs included in an analytical sequence was defined as half the
the retention time window. For chlordane, ±0.10 and ±0.05 min were used
for the retention time window . for the OV-17 and DB-5 columnsi
respectively.
The external standard quantitation method was used taking th*
response factors from the standard runs at the beginning of a sequence*
However, the GC/ECD response was ensured through a sequence by monitoring
the individual standard mixtures run at the middle and the end of *
sequence. Except for pentachlorophenol and dicofol, only in a few cases
(<0.5Z) did the response factors of the targeted compounds from these
individual standard mixtures differ over 25Z from those of the initial-
standard mixture. Pentachlorophenol and dicofol usually showed large
deviations. The quantitation value from the OV-17 column was used
data reporting unless an Interference occurred.
36
-------
GC/MS/MID Analysis
GC/MS/MID analysis was used for the identification and quantitation
of the GC/MS/MID target compounds (Table I) and secondary confirmation for
all GC/ECD target compounds except chlordane. A Finnigan 4500 mass
spectrometer with a Finnigan 9611 GC was used for analysis with a 30-m X
0.25-mm ID DB-5 bonded fused silica analytical column. The GG was set at
50 C for 2 minutes and programmed to 295*0 at 10°C/min. The GC injector
was 235*C and injection volume 2 n\. The electron energy was set at 70 eV
and scan time of 1 scan/s. Daily mass calibration and tuning were
performed with perfluorotributylatnine (FC-43). Five selected ion
descriptors listed below were used. The time interval for each mass in
any of these descriptors was set at 0.05 s. The chromatogram of a
standard injected at 2 ng/^1 with the ranges of the five descriptors
indicated is shown in Figure 1.
Five GC/MS/MID Descriptors
SI :m/e 109,110,115,126,141,151,152,166,170,185
52 :m/e 109,Illtl37,165,173,179,183,188,199,200,219,266,284
S3 tm/e 97,100,115,125,144,173,197,261,263,272,285,301,332
S4 :m/e 57,81,115,119,130,149,185*220,246,260,263,265,318,320,353
» sn/e 123,139,143,152,165,171,183,227,235,237,240,251,263,265
The internal standard quantitation method utilizing two internal
standards, DlO-phenanthrene and D12-chrysene, both at 2 ng/pl, was used
for GC/MS/MID analysis. The instrument was initially calibrated with a
five-point curve using the concentrations 0,4, 2, 4, 6 and 8 ng/j*l. The
relative standard deviations for the response factors of all GC/MS/MID
target compounds was required to be less than 302 before sample analysis.
Quantitation was performed using the response factor from a continuing
CM brati°n standard at 2 ng//*l injected at the beginning of every 12-hour
shift. Samples were then analyzed provided the difference between the
response factor from the continuing calibration standard for each
GC/MS/MID target compound and the average response factor from the five-
point calibration curve did not exceed 25Z. All the analyses, including
GC/ECD and GC/MS/MID, were to be completed within 30 days after sample
extraction.
Results
The surrogate recoveries and matrix spike results obtained during the
course of the study were considered satisfactory.3 However, three
Problems did occur. The first problem was caused by the boiling chips
used for extraction.3 For an unknown reason, some pesticides were removed
*>y the Boileezer® chips used. Therefore, the boiling chip used for the
e*traction process must be demonstrated to be neutral to the pesticides
studied. The second problem arose from the instability of certain target
compounds, such as captan, chlorothalonil, folpet, oxychlordane, propoxur
*tvd trans-perraethrin, which were found to degrade in the standard solution
sven at 4*C. The standard solutions had to be prepared every three months
p °m neat materials. The third problem was of a chromatographic nature,
V^ntachlorophenol and dlcofol, which usually showed quite a. large
Aviation in response factors from one standard to another standard, ware
Poorly chromatographable compounds due to their high polarity.-* Dicofol
Was also partially thermally decomposed to 4,4*-dichlorobenzophenone,
Specially in the DB-5 jnegabore column. Therefore, 4,4'-
37
-------
dichlorobenzophenone was also monitored when using the DB-5 megabon
column for GC/ECD analysis.
Conclusion
Stringent criteria have been followed in daily operations to analyze
more than one thousand samples with little effort required for repeated
analyses. The analysis methods gave quite satisfactory results for all
the analytes other than pentachlorophenol and dicofol.
References
1. R. G. Lewis and K. E. MacLeod, "Portable sampler for pesticides ani
semivolatile industrial organic chemicals in air," Anal. Chem. J54(2):
310-5 (1982).
2. "Method 608--Organochlorine pesticides and PCBs," Federal Registei,
42: 209, pp. 89-104 (Oct 26, 1984).
3. J.P. Hsu, H. G. Wheeler, Jr., D.E. Camann, H.J. Schattenberg III,
R,G. Lewis, and A.E. Bond, "Analytical methods for detection of
nonoccupational exposure to pesticides," J. Chromatog. Sci. 26: 181-!
(1988).
38
-------
TABLE I, QUARTITATION LIMIT GOALS OF TARGET PESTICIDES
Quantitation Limit Go
Comound
n on <.mm
ft
GC/ECD
chlordane
Individual Standard Mixture I
7-BHC
ronnel
chlorpyrifos
captan
4,4'-DDD
4,4'-DDT
methoxychlor
cis-permethrin
Individual Standard Mixture II
dtchlorvos
heptachlor
oxychlordane
dieldrin
2,4-D butoxy ethyl ester
trans-permethrin
Standard Mixture IV
pentachlorophenol
dicofol
GC/MS/MID
ortho-phenylphenol
propoxur
bendiocarb
atrasine
diazinon
carbaryl
raa lathi on
resmethrin
0 . 8
0.05
0.07
0.06
0.3
0.06
0.06
0.1
0.4
2
0.07
0.06
0.08
1
0.4
Individual Standard Mixture III
hexachlor obenz ene 0.03
o-BHC 0.04
chlorothalonil 0.04
aldrin 0.05
dacthal 0.05
heptachlor epoxide 0.04
4,4'-DD£ 0.06
folpet 0.2
4
1
0.4
0.2
0.5
0.5
0.6
0.5
0.5
ng/m *
145
9
13
11
55
11
11
18
73
364
13
11
15
182
73
5
7
7
9
9
7
11
36
727
182
36
18
45
45
55
45
45
91
Based on the assumption that extracts were concentrated to 10 ml,
8PHt to two 5-ml portions, concentrated to 1.0 ml and 0.50 ml for
GC/ECD and GC/MS/MID, respectively, and 2 /il was injected to either
Instrument. The total volume of air sampled was assumed to be 5.5 m^.
39
-------
189.
RIC _
}-
—
—
_
ILCf
D
600
41
S1 S2
*
875
,l
5
800 IE
• .j
1
1)0
S3
11
35
i
S4
S5
13
13
rr^
T1
IB
I
1200E
56
1463
A
15 4
1400 1606
371208,
Scan
Figure 1. GC/MS/MID calibration standard at 2 ng/yl
-------
A SYSTEMATIC PROCEDURE FOR CHLORDANE IDENTIFICATION IN AIR
Herbert J. Schattenberg, III
David E. Camann (presenter)
Southwest Research Institute
San Antonio, Texas
Chlordane, a multicomponent termiticide, was one of 33 pesticides
monitored in air in the U.S. Environmental Protection Agency's
Nonoccupational Pesticide Exposure Study (NOPES). The chromatographic
peak pattern of technical chlordane can be significantly altered by
environmental degradation, especially at the low concentrations
encountered in air samples. Thus, a systematic procedure is preferable to
the usual fingerprint identification approach for multicomponent
chromatographic analysis.
A systematic procedure based on analysis by electron capture gas
chromatography using two dissimilar megabore capillary columns was
developed for the NOPES program. The procedure utilized ten major peaks
on each column as the chromatographic identifiers of chlordane. ¥hile
permitting judgment of the experienced analyst in some areas, the protocol
employed strict criteria regarding presence of major peaks and relative
component responses. The procedure was found to give consistent,
reliable, and largely unbiased sample analyses. The median difference of
collocated chlordane air measurements was
42
-------
Introduction
Technical chlordane is a mixture of at least 50 components produced
by the Diels-Alder addition of cyc1 opentadiene to
hexachlorocyclopentadiene. Until 1983, it was used extensively in
treating the soil beneath residences to prevent and control termite
damage. Chlordane was one of 33 pesticides recently monitored in
approximately 260 different homes by the U.S. Environmental Protection
Agency in its Nonoccupational Pesticide Exposure Study (HOPES). The
purpose of the NOPES program was to test methodology and survey
residential exposure to common household pesticides from potential
inhalation, ingestlon of food and water, and acute dermal contact during
application.
At trace environmental levels, technical chlordane is usually
analyzed by gas chromatography and subjectively identified by recognizing
the chromatographic peak pattern (see Figure 1) . Unequivocal chlordane
identification in a sample based on "fingerprint* overlay comparison of
the numerous peaks is obtained only in ideal situations. Difficulties
occur, especially at low concentrations, because the peak pattern can be
significantly altered by environmental degradation ("weathering") and by
selective adsorption on the trapping medium. In an indoor air sample
taken several years after application, the relative concentrations of the
individual components may differ markedly from their composition in the
original technical chlordane mixture. Differences in component
volatility may be a major factor.
To overcome these problems, a systematic procedure and a detailed
protocol were developed as a part of NOPES for chlordane identification
and quantification in low level environmental samples. This paper
presents and characterizes the chlordane procedure.
Experimental
Indoor, outdoor, and personal air samples, drinking water, and gloves
worn during pesticide application were taken for analysis. The air
samples were collected using the method of Lewis and MacLeod2 by drawing
approx 5.5 ra^ of air through a polyurethane foam (PUF) plug over a 24-h
period with a Du Pont Alpha-1 constant-flow sampling pump. The PUF and
glove samples were Soxhlet-extracted. The final extracts were split for
analysis for chlordane and 24 single-component pesticides by gas
chromatography (GC) with electron capture detection (GC/ECD) and for all
33 target pesticides except chlordane by GC/mass spectroscopy (GC/MS), as
described by Hsu, et al.^
The GC/ECD analysis was performed on a Hewlett Packard (HP) 5890
chroroatograph equipped with two dissimilar megabore (0.53 mm ID) capillary
columns, two ECDs, and two HP 7673 autosamplers. The liquid phases were
OV-17 (50% phenylmethylsilicone, Quadrex Corp., New Haven, CT) for the 15
m primary column and. DB-5 (5% phenylmethylsilicone, J&W Scientific,
Folsom, CA) for the 30 m confirmation column.
Description of Method
Technical chlordane was identified using an explicit step-by-step
protocol based on the ten raost intense peaks observed on each column.
These Included four named components: a-chlordane (cis), f-chlordane
(trans), trans -nonachlor and heptachlor. Figure 1 shows typical
43
-------
chromatograms of a chlordane standard on both columns. The first threi
columns of Table I present retention time (RT) data for the ten selected
peaks. It should be noted that the elution order of trans-nonachlor ant
a-chlordane on the OV-17 column is reversed from that observed on the DB-!
column.
When an unknown sample was analyzed, the resulting RT and response
data for the ten selected peaks in the standard and the sample wen
tabulated for both columns, as shown in Table I. A retention window o!
±0.10 min about each standard peak absolute RT was used on the primary OV-
17 column to select sample peaks.
The flowchart in Figure 2 shows the protocol followed to providt
consistent interpretation guidelines to the analyst in making a tentativf
assignment of chlordane based on the sample peaks within the ±0.10 windot
on the primary column. The absence of either of the two mail
stereoisomers, a- or 7-chlordane, was cause to report the sample negativf
for chlordane. When these two peaks were large, within a factor of fivf
of each other, and at least two of the other eight selected peaks wen
also present in the sample, then a tentative identification of chlordam
was made. If all or most of the major peaks of the selected ten wen
present, as specified in Figure 2, then a tentative assignment was mad-
without resorting to comparison of the areas of the a- and 7-chlordane
peaks.
Once a tentative assignment of chlordane had been made using the OV-
17 column, results from the DB-5 column were evaluated as the confirmatioi
procedure. The RT of those peaks present in sample were compared witt
those from the standard, using a more restrictive retention window of
±0.05 min about each standard reference peak. As indicated in the example
shown in Table I, only the four individual peaks identified on botl
columns could unequivocally be confirmed or rejected by the confirmatioi
analysis. If either the a or 7 isomer were absent from the confirmatiot
analysis, the sample was considered negative for chlordane. If either 01
both of the two other known components, trans-nonachlor and heptachlor,
were absent in the confirmation analysis, the corresponding peak was
deleted from the primary analysis, and tentative assignment was re-
evaluated as shown in Figure 2. A minimum of two peaks besides the a ant
7 isomers must have been present in the confirmation analysis to confiic
the presence of chlordane.
Situations have been observed in which a chromatographic shift caused
many of the sample peaks to fall slightly outside the specified retentioi
windows. The shifts, which may be produced by temperature or carrier floi
fluctuations, became obvious when seen in several samples analyzed in tht
same sequence. The protocol allowed an experienced analyst to move the
windows in the direction of the shift. In this study possible shifts or
the confirmation column were detected by screening the sample peaks usinj
±0.10 min windows in addition to the specified ±0.05 min windows.
Quantitation was performed using an external standard technique after
the sample had been identified as positive for chlordane. The peak areas
for all components positively identified on the primary analysis which had
not been rejected during the confirmation analysis were summed to give the
total peak area for the sample, as shown in Table I. The areas of the
corresponding peaks in the standard were summed to give the total area of
the standard. Summations for quantitation purposely excluded the
component heptachlor since that compound is used independently as a
pesticide and is commonly added to termiticide formulations containing
44
-------
technical chlordane.
from equation (l)i
The chlordane concentration in air was determined
Chlordane
Concentration
(ng/m3)
Sum of Sample
Peak Areas*
Amount
of Standard
Injected (ng)
x Extract
Volume (/*!)
Sum of Same
Standard Peak
Areas*
Sample Injection
Volume
-------
Conclusions
A systematic ten-peak approach has been developed for both the
identification and quantitation of technical chlordane in weathered
environmental samples. This procedure provides guidelines for analyst
interpretation as an alternative to pattern recognition of multicomponeoti
mixtures. The accuracy and precision of the air measurements show that
the procedure has produced consistent and reliable results in analyzing
numerous samples.
References
1. G.W. Sovocool. R.G. Lewis, R.L. Harlen, N.K. Wilson, R.D. Zehr
"Analysis of Technical Chlordane by Gas Chromatography/Masf
Spectrometry," Anal. Ch«»m. AQi 734-40 (1977).
2. R. G. Lewis, K. E. MacLeod, "A portable sampler for pesticides arv4
semivolatile industrial organic chemicals,* Anal. Chem. *&: 310-5
(1982).
3. J.P. Hsu, H.G. Wheeler, D.E. Camann, H.J. Schattenberg, R.G. Lewis
A.E. Bond, "Analytical methods for detection of nonoccupational
exposure to pesticides," J. Chromatogr. Sci. ?6i 181-9. (1988).
46
-------
TABLE I. TYPICAL CHLORDANE IDENTIFICATION AND QUANTITATION WORKSHEET
Standard Sample
Paal
L Compound
RT . min
Area
HT. mln
Area
Primary (OV-17> nnlnmn
1
2
3
4
5
6
7
8
9
10
Sum
Coni
1
2
3
4
5
6
7
8
9
10
Heptachlor
7-chlordane
t-nonachlor
a-chlordane
12.74
18.07
18.80
21.17
22.75
23.64
23.87
24.32
27.51
28.36
excluding heptachlor
Heptachlor
7-chlordane
a-chlordane
t-nonachlor
16.56
22.23
23.19
24.98
26.37
27.97
28.72
28.99
31.97
32.40
58,000*
266,000
328,000
215,000
130,000
605,000
261,000
599,000
116,000*
137,000*
2,076,000
28,000
100,000
143,000
46,000
87,000
260,000
247,000
167,000
57,000
72,000
18.02
18.72b
21.11
22.75
23.61b
23.88b
24.29b
23.17b
26.35
27.95b
28.70b
28.97b
20,000
128,000
44,000
9,000
26,000
15,000
23,000
137,000
14,000
4,000
8,000
6,000
5,000
a Peak excluded from area summation since absent from sample,
b Primary peak confirmed on DB-5 column.
47
-------
TABLE II. CHLORDANE ANALYTICAL ACCURACY FROM
ANALYSIS OF BLIND SPIKED PUF SAMPLES
Percent Bias
-16
+6
-15
a PUF samples prepared by external QA laboratory; each was
spiked with 6 or 7 additional pesticides.
b Not detected at stated quantitation limit goal.
Sample5
N6
N7
N8
N9
Spiked
fng/sample )
2389
3201
None
4777
Reported
fug/sample )
1997
3395
<800b
4058
TABLE III. CHLORDANE MEASUREMENT PRECISION FROM
ANALYSIS OF COLLOCATED FIXED AIR SAMPLES
Chlordane Concentration in Air^
Indoor Outdoor
Primary Duplicate Primary Duplicate
Set 1 76 76
258 245 (211)b <154C 45 (43)b
Set 2
49 59
121 148
89
49
2740
41
<152C
66
289
182
103
70
<149C
142
58
59
2630
<124C
67
65
257
<154C
93
36
64
125
<121C 41
Set 3 182 <154C 75 67
43 40
<147C 25
a Nominal quantitation limit goal of 146 ng/m3 when 5.5 m3 of air
was sampled.
b Collocated triplicate sampler result.
c Not detected at stated quantitation limit goal.
48
-------
(a) OV-17 Megabore
Column
LJL-iJ
DB-S Megaboxe
Column
Figure 1. Typical chromatograms of technical chlordane on
(a) OV-17 and (b) DB-5 columns
49
-------
PEAKS 6 OR 8
NOT PRESENT
T
SAMPLE NEGATIVE
FOR CHLORDANE
COMPARE RT OF 10
PEAKS. ALL 10
WITHIN RT WINDOWS
(±0.10 OF RT)7
#2,3,6,7,8,9,4
10 IN WINDOWS?
NO
12,3,6,8,9, fc
10 IN WINDOWS?
NO
#3,6,8,9,& 10
IN WINDOWS?
NO
ALL 10
(MINIMUM ACCEPTABLE FOR
CHLORDANE ID) *6, 8 IN
WINDOWS AND AT LEAST 2
OTHER PEAKS (OF 1, 2, 3,
4,5,7,9 & 10) IN WINDOWS?
TENTATIVE ASSIGNMENT
BEGIN CONFIRMATION
STEP
YES
YES
YES
YES
INSPECT AREA
RATIOS OF 6
VS 8 (RATIOS
WITHIN FACTOR
OF 5)7
NO
YES
NO
SAMPLE NEGATIVE
FOR CHLORDANX
Figure 2. Flowchart for tentative identification of chlordane
using the primary column
50
-------
Chlorinated Pesticides and Polychlorinated Biphenyls
in the Atmosphere of the Canadian Arctic
G.W. Fatten, D.A. Hinckley, M.D. Walla, and T.F. Bidleman
Department of chemistry and Marine Science Program
University of South Carolina, Columbia, South Carolina, 29208, USA
B.T. Hargrave
Marine Ecology Laboratory, Bedford Institute of Oceanograpy
Dartmouth, Nova Scotia/ B2Y4A2, Canada
INTRODUCTION
The Canadian Arctic is a region with low population density and
limited sources of local pollution. Despite the archipelagos' remote-
ness, the Arctic air mass is polluted by anthropogenic emissions in
temperate latitudes. Long-range transport of pollutants has been
observed since the Arctic haze aerosol was reported in 1956. Eastern
Europe and Asia are the most likely sources of the aerosol pollu-
tion. (1) organochlorine (OC) pesticides and industrial chemicals such
as polychlorinated biphenyls (PCBs) and hexachlorobenzene (HCB) have
oeen reported in Arctic air. (2, 3) Hexachlorocyclohexane (HCH), diel-
arin, and DDT were measured in the snow of east-central Ellesmere
Island. (4)
Atmospheric deposition is a likely source of OC to the Arctic
ecosystem. OC pesticides and industrial chemicals (HCB and PCB) have
°een observed in Arctic marine organisms since the early 1970 's.( 5/6)
These pollutants quickly move through the uncomplicated polar food
chains and can reach high levels in Arctic animals. A better under-
standing of OC contaminants in this region is important because of the
fragile ecosystem and the native people who rely on marine animals for
Pjeir food supply. In 1986, Canadian groups from Bedford Institute of
oceanography and Arctic Laboratories initiated an investigation of OC
input to and transfer through Arctic food chains. (7) As part of this
pro]ect, we determined OC pesticides and PCB in air and surface water
m the high Arctic,
METHODS
In 1984, the Canadian Polar Continental Shelf Project established
J research camp on a tabular iceberg that calved off the Ward Hunt Ice
jneif. The 7 x 4 km x 45 m thick ice island is presently located off
northwest shore of Axel Heiberg island (Figure 1). The ice island
provides a safe sampling platform in a region where shipboard work is
not practical due to extensive ice cover. Air samples were taken
August-September, 1986 and air, snow, and seawater samples were taken
1987.
Air volumes of 1400-3000 m were pulled through a glass fiber
J-Uter (GFF) and two polyurethane foam (PUF) plugs at flow rates of
u-6 - 0.8 cu'm/min.(8) Collections were made at least 2 km from the
ca™P and electrical power was supplied by gasoline generators operated
-30-100 m downwind of the samplers. Four seawater samples were taken
at a depth of 10 m using a National Bureau of Standards sampler. One
51
-------
sample was taken at 1 m depth by plunging a bottle into a lead of
recently exposed water. The salinity was 32 parts-per-thousand and
the water temperature was -1.7 C. Seawater (2-3 L) was pulled at 0.8
-1.3 L/h through a GFF followed by front and back 500-mg C-8 cart-
ridges (J.T. Baker SPE).(9)
PUF plugs were extracted in a Soxhlet apparatus with petroleum
ether. GFF (air samples) were refluxed with dichloromethane. GFF and
PUF extracts were cleaned up and fractionated using an alumina-silicic
acid column chromatography procedure.(8) The C-8 cartridges were
eluted with 3 mL of 1:1 ethyl ether/hexane. GFF (water samples) were
refluxed with acetone for 4 hr. Analyses of air and water samples
were carried out using GC-ECD. Verification of chlordane components,
polychlorocamphenes (PCC), DDE, and DDT was carried out on air samples
using capillary GC-mass spectrometry in the negative ionization mode.
Of the compounds collected, pentachlorobenzene (PeCB) and HCB were the
only ones that showed breakthrough to the back PUF plugs. PeCB was
found in equal amounts in both traps and no quantitative results could
be obtained. HCB showed breakthrough ranging from 27-99% of the front
trap value (average » 53%) and HCB results were calculated by summing
the quantities on both plugs; for some samples this may represent a
lower limit.
Air Results
Concentrations of OC found on PUF traps for the August 1986 and
June 1987 trips are given in Table 1. GFF were analyzed for a few
samples, but none showed detectable OC residues. HCB, o-HCH, and
y-HCH' were found in all samples at concentrations similar to most
values reported in the northern hemisphere. Alpha-HCH levels were the
highest of all the OC. The ratio of mean concentrations for <*-HCH/r-
HCH were 17.6 for August and 7.6 for June. Researchers in the Nor-
wegian Arctic also found that the o-HCH/r-HCH ratio was lower in the
winter-spring than in the summer-fall, and suggested that the propor-
tion of o-HCH is higher in older air masses because of photochemical
transformation of r-HCH to «-HCH.(2) However the a-HCH/V-HCH ratio
may also be influenced by different rates of atmospheric deposition of
the isomers and by transport of HCH from regions where lindane or
technical HCH are mainly applied.
The sum of four chlordane components are given in Table 1. The
August mean is close to the total of the same four chlordanes at Mould
Bay in June, 1984, but our June mean is nearly three times higher.(3)
Average concentrations of cis-chlordane (CC) at the Ice Island were
also higher than those observed in the Norwegian Arctic.(2) Average
ratios of the chlordanes (R - trans-chlordane (TC)/cis-chlordane) at
the Ice Island were: August R » 0.39; June R » 0.60. This order of
abundance was also observed by Hoff and Chan at Mould Bay where the
mean R in June, 1984 = 0.56. Expectations are for TC to be the
dominant isomer in air, since it is the most abundant and volatile of
the chlordanes. Reported TC/CC ratios from more temperate areas were
greater than unity (Figure 2). The reason for the depletion of TC in
Arctic air is unknown.
Concentrations of PCC were similar for both expeditions. Traces
52
-------
of PCC have been previously identified in the Norwegian Arctic, but
were not quantified. No other information is available about the
presence of PCC in Arctic air, although PCC have been found in ambient
air from more temperate latitudes.(11) Of the identified OC in Arctic
air, PCC are third in abundance.
PCB were calculated as Aroclors 1242 or 1254 using response
factors derived from the whole Aroclor chromatographic patterns, and
also by summing individual PCB congeners. Average total PCB' concen-
trations derived by these two methods agreed within 14%.
Two samples showing clear DDE peaks by GC-ECD were examined by
®Jectron impact GC-MS, and DDE was confirmed in both. However only
the highest-volume air sample yielded a positive DDT result by GC-MS,
even though GC-ECD chromatograms of other samples showed an apparent
DDT peak.- Co-eluting PCC may have given a false positive for DDT by
GC-ECD, and therefore the DDT concentrations in Table 1 should be
considered upper limits.
Eleven air samples were analyzed for chlordanes and PCC by GC-
negative ion mass spectrometry (GC-NIMS). Two chlordanes (TC and CC),
trans-nonachlor (TN), cis-nonachlor (CN), and PCC components contain-
ing 6-9 chlorines were clearly identified on front PUF plugs; back
plugs were clean. The most volatile constituents of each PCC homolog
dominated the chromatograms. Chlordanes and PCC in the samples were
quantified by GC-ECD and by GC-NIMS. A comparison of the. two techni-
ques is given in Figure 3. Perhaps ECD results for TC and TN were
slightly inflated by interferences underlying these peaks. However,
considering the analytical difficulties at these ultra-trace levels,
we feel the agreement between GC-ECD and GC-NIMS results is
satisfactory.
Water/Snow Results
Because of the small water volumes sampled, HCH were the only OC
M utet* in seawatei:- Average concentrations are given in Figure 4.
NO breakthrough to the back cartridge was observed for any of the
samples nor were HCH found on GFF. Mean air and water concentrations
iron the June trip were used to estimated the state of air-water
equilibrium for the HCH isomers. The Henry's Law constants (H) for
tne HCH isomers were taken from the literature and corrected for
salinity and temperature using a method described in an earlier
Paper.(13) Figure 4 gives the calculated equilibrium concentrations
°t HCH in surface water, and actual concentrations. Alpha-HCH appears
J? J8 close to equilibrium, but y-HCH is undersaturated. Whether the
difference between the two isomers is real or an artifact of
uncertainties in H is unknown. A likely source of error in these
Si?113*110118 is t*16 extrapolation of H from 25°C to -2°C. A
fetter knowledge of H is needed, especially as a function of
temperature.
53
-------
ACKNOWLEDGEMENTS
This work was supported by NATO Scientific Affairs Division
(Grant No. NATO 04-0667-86), the University of South Carolina Venture
Fund, and the Canadian Polar Continental Shelf Project. Thanks to
Bedford Institute of Oceanography and Arctic Labs for extending the
invitation for us to participate in this project, and for their help
with the field sampling. A special thanks to the excellent PCSP staff
on the Ice Island.
REFERENCES
1. Barrie, L. A. 1986. Atmos. Environment, 20, 4, 643-663.
2. Pacyna, J. M., Oehme, M. 1988. Atmos. Environ., 22, 2, 243-257.
3. Hoff, R. M., Chan, K.-W. 1986. Chemosphere, 15, 449-452.
4. McNeely, R., Gummer, W. D. 1984. Arctic, 37, 210-223.
5. Bowes, G. W., Jonkel, C. J. 1975. Canad. J. Fish. Aquat. Sci.,
32, 2111-2123.
6. Norstrom, R. J., Muir, D. C. G. 1987. Toxic Contamination in
Large Lakes, ed. N. W. Schmidtke, Lewis Pub., Vol. 1, 83-112.
7. Hargrave, B. T., Vass, W. P., Erickson, P. E., Fowler, B. R.
1987. Tellus B, in press.
8. Billings, W. N., Bidleman, T. F. 1983. Atmos. Environ., n,
383-391.
9. Hinckley, D. A., Bidleman, T. F. 1987. Pittsburgh. Conf. Abst.,
No. 109.
10. Bidleman, T. F., 1988. unpublished data.
11. Bidleman, T. F., Wideqvist, U., Jansson, B. ,S6derlund, R. 1987.
Atmos. Environ., 21, 641-654.
12. Foreman, W. T., Bidleman, T. F., 1987. Environ. Sci. Technol.,
21, 869-875.
13. Patton, G. W., Hinckley, D. A., Walla, M. D., Bidleman, T. F.,
Hargrave, B. T., 1988. Tellus B, in press.
54
-------
ARCTIC OCEAN
61°N. 100° W
FIGURE 1. LOCATION OF THE ICE ISLAND DURING AUGUST-SEPTEMBER,
1986 AND JUNE 1987.
Compound
Mean pg/m
Summer 1986 Summer 1987
a-HCH
Y-HCH
HCB
PCC
PCBa
Cis-Chlordane
Trans-Chlordane
Cis-nonachlor
Trans-nonachlor
P.p'-DDE
P,p'-DDT
546
31
189
44
14
2.8
1.1
0.4
1.5
0.1
0.9
340
45
147
36
20
4.
2.
0.
3.
2.
2.
0
3
7
7
9
3
a- PCBs calculated as sum of Aroclor 1242 + Aroclor 1254.
Table 1. ATMOSPHERIC ORGANOCHLORINES AT THE ICE ISLAND.
55
-------
2.0
1.8
1.6
z! i*
••'
o
0.8
0.6
0.4
0.2
0.0
ICE MOULD SflBLE SWEDEN COLUMBIfl CHLOR
ISLRND BRY ISLRND SC VRPOR
(3) (10) (11) (12)
FIGURE 2. TRANS-CHLORDANE / CIS-CHLORDANE RATIOS IN AIR
100r
>-
I- 80
=> 60
O
ill
>
K 40
LU
cc
20
o
o
TC
CC
TN
PCC
EflUIL. fCTUflL
alpha-HCH
EOUIL. FCTURL
gamma-HCH
FIGURE 3. COMPARISON OF GC-ECD TO GC-NIMS.
FIGURE 4. HCH CONCENTRATIONS IN ARCTIC WATER AND PREDICTED AIR'
SEA EQUILIBRIUM VALUES.
56
-------
METHOXYLATED PHENOLS AS CANDIDATE
TRACERS FOR ATMOSPHERIC WOOD
SMOKE POLLUTION
Steven B. Hawthorne, Mark S. Krieger, and David J. Miller
University of North Dakota Energy and Mineral Research Center
Grand Forks, North Dakota 58202
Robert M. Barkley
Cooperative Institute for Research in Environmental Sciences
University of Colorado, Boulder, Colorado 80309
Unfractionated extracts of wood smoke particulates have been analyzed using a
combination of capillary gas chromatography coupled with low resolution mass
spectrometry (GC/MS), GC coupled with high resolution mass spectrometry (GC/HRMS),
and chemical ionization mass spectrometry with deuterated methanol as the reagent
gas. Although several PAHs, oxy-PAHs, and phenols were identified in the extracts,
thirty of the most concentrated species were derivatives of guaiacol (2-
lethoxyphenol) and syringol (2,6-dimethoxyphenol). Samples collected from smoke
lumes cooled to near ambient temperature onto filters backed up by polyurethane
oam (Pup) plugs showed that some of the methoxylated phenols were primarily in the
'aP°r Phase, while some were primarily associated with the particulates. Hardwood
pine smoke showed similar concentrations of guaiacol derivatives, but the
concentration of syringol derivatives was much higher in hardwood smoke. Since the
uaiacol and syringol derivatives are pyrolysis products of wood lignin, they are
-xpected to be unique to wood smoke in urban atmospheres and are therefore
uggested as tracers for atmospheric wood smoke pollution. The collection,
entification, and quantitation of these candidate tracers from several
:esidential wood stoves is reported.
57
-------
Introduction
The use of residential wood-burning appliances contributes 30 to 80% of &
winter urban air fine participate loading in several communities, and has bee
estimated to account for more emissions of polycyclic organic compounds than an-
other source1. Studies of the relative impact of wood smoke, vehicle exhaust, ac
other particulate sources are presently limited by a lack of tracer species uniqu
to wood smoke particulates. Chemical tracers such as methylchloride and potassh
have been applied, but their usefulness has been limited by high and variabti
background levels (for methylchloride) and by highly variable concentrations o;
potassixnn on wood smoke particulates. The use of potassium as a tracer is furthe:
complicated by its presence in fine particulates from soil. Organic tracerr
including retene (l-methyl-7-isopropylphenanthrene) and levoglucosan (the anhydride
of -glucose) have also bean suggested, but their use has been limited. Tfe
Tneasurement of i(*C can be used to distinguish between "new" carbon (from woo!
burning) and "old" carbon (from fossil fuel combustion), but the analysis takes t
relatively long time to perform and requires instrumentation that is not wideb
available.
Investigations that have been conducted into the identification of organics
extracted from wood smoke particulates have focused on fractions containing
polycyclic aromatic hydrocarbons (PABs) and oxy-PAHs, while much less emphasis has
been placed on the more polar fractions. We have identified approxiaiately 30
methoxylated phenolic species in unfractionated extracts from wood smoke
particulates and from polyurethane foam (PDF) plugs used to collect vapor phase
organics from smoke plumes. These candidate tracer species were identified using
capxllary GC/MS with electron impact (El) ionization, GC coupled with higl
resolution mass spectrometry (GC/HRMS), and chemical ionization (CI) mass
spectrometry with deuterated methanol as the reagent gas. Twelve of the most
concentrated species were quantitated in particulate and PUP samples collected in
the smoke plume from six different residential wood stoves.
Experimental Methods
Wood smoke particulate samples were collected from six different residential air-
tight wood stove installations, three of which were burning a mixture of hardwood
species (primarily oak and ash) and three of which were burning pine. Samples were
collected at 4 L/min onto 37 ton glass fiber filters backed up by two 37 mm diameter
X 40 ma long polyurethane foam (PUF) sorbent plugs. Prior to sample collection,
the PUF plugs were pre-extracted for four hours using several changes of acetone
with sonication, dried under clean air, and stored in brown glass bottles with
teflon-lined caps until used. After the wood smoke samples were collected, each
filter and PUF plug was extracted for two hours using sonication with 60 mL of
acetone which contained 1 jig of d^guaiacol as an internal standard. Following
extraction, the samples were evaporated under nitrogen to 1 mL. A 1-gram sample of
hardwood soot scrapings was extracted in a similar manner in order to yield a more
concentrated sample for GC/HKMS analysis.
All low resolution GC/MS (El and CI) analyses were performed using a Hewlett-
Packard model 5985B GC/MS equipped with a dual El/CI source. Chromatographic
separations were achieved using a 20 m X 0.25 on i.d. (0.25^m film thickness) DB-5
capillary gas Chromatographic column (J & W Scientific). El mass spectra were
obtained at 70 eV with a typical scan range of 60 to 350 amu. Gas chromatography
with flame ionization detection (GC/FID) was performed using the same
Chromatographic conditions with a Hewlett-Packard model 5890 GC.
Chemical ioniaation mass spectrometry2 was performed using methanol-dj (CH3OD) as
the reagent gas at a source pressure of 0.2 torr. CI tuning parameters were
58
-------
optimized by maximizing the intensity of the M+3 ion (m/z 125) of 2,4-
dimethylphenol introduced through the direct insertion probe. Because of a
prominent background ion at m/z 101 resulting from the reagent ion cluster
(01300)30+, the scan range when using CI was 105 to 350 amu.
Capillary gas chromatography/high resolution mass spectrometry (GC/HRMS) was
performed using a VG 7070 EQ-HF equipped with a Hewlett-Packard model 5890 gas
chromatograph. Chromatographic conditions were similar to those described above.
The mass spectrometer scan range was 50 to 400 at 1.0 sec/decade and a resolution
of approximately 4000 (10% valley).
Results
Figure 1 shows a typical chromatogram obtained from the GC/FID analysis of the
unfractionated extract of hardwood smoke soot. Preliminary GC/MS analyses using El
lonization showed that the all of the hardwood and pine smoke extracts contained a
complex mixture of phenols, oxygenated phenols, PAHs, and oxy-PAHs. The combined
use of GC/MS, GC/MS with CH3OD CI, and GC/HRMS allowed the identification of
approximately 55 individual species shown on Figure 1. Several phenols, PAHs, and
oxy-PAHs were identified based on a comparison of their El mass spectra with
standard spectra, and in several cases by a comparison of their retention indices
with those of known compounds. Many of these species including the PAHs and oxy-
AHs are typical products of combustion processes, while several of the species
^•e., the phenols) would be expected from the pyrolysis of wood lignin.
The guaiacol (2-methoxyphenol) and syringol (2,6-dimethoxyphenol) derivatives
were more difficult to identify. GC/HRMS analysis was used to determine the
molecular formula of each these species, however, the HRMS data and the El
fragmentation patterns were not sufficient to differentiate among the several
•Afferent structural isomers that were possible for a particular molecular formula.
Tftis problem was particularly severe when trying to determine the number of -OH and
R functionalities present on a particular species (e.g., a dimethoxybenzene
versus a methylmethoxyphenol versus a dihydroxydimethylbenzene) since the El mass
spectra of such isomers are frequently indistinguishable from one another.
In order to determine the type of oxygen functionalities present on the
oxygenated phenols , chemical ionization GC/MS analysis using CH30D as the reagent
gas was used to count the number of -OH groups present on each component2. Species
with no -OH group show a base peak (100% relative intensity) at M+2 from the
'H!T *0tl °f D+* In Contra8t» species with one -OH show a base peak at M+3 from the
addition of D+ and the exchange of one -OH for -OD, while species with two -OH
groups show a base peak at M+4 from the addition of D+ and the exchange of two -OH
ydrogens to form two -OD groups. For example, dimethoxybenzene,
ethylmethoxyphenol, and dihydroxydimethylbenzene isomers all have a molecular
weight of 138, but would show base peaks using CHsOD CI at m/z 140, 141, and 142,
respectively.
The combined use of GC/MS with C^OD CI, GC/HRMS, and El fragmentation patterns
allowed 30 oxygenated phenols to be identified in the wood smoke particulate
extracts as shown in Figure 1. The identities of fifteen of the oxygenated phenols
were also confirmed by a comparison of retention indices and El fragmentation
Patterns with those of known standards. Comparisons of retention indices showed
"at substituent groups on sample species were para to the -OH group. In the cases
ere standards were not available, the substitution of the substituent group
•f*8-» methylsyringol) would also be expected to be para to the -OH group based on
«»e structure of wood lignin.
In order to determine the types and quantities of individual raethoxylated phenols
59
-------
emitted from different wood stoves and from hardwood versus pine, samples i
particulate and vapor-phase organics were collected from six different residentii
wood stoves as described above. During collection of the wood smoke samples ti
ambient temperatures ranged from -2 to -6 °C, the temperature of the smoke at th
chimney outlets ranged from 40 to 80 °C, and the temperature of the smoke at tk
collection point (1/2 m from the chimney outlet) was approximately 5 °C abor
ambient. Following sample collection, the samples were extracted as previousl
described. Extracts of the backup PUF plugs showed no significant specie
demonstrating that the vapor-phase organics were efficiently collected on the firs
PUF plug. Extracts of filter and PUF blanks, and the second extracts of the sampl
filter and PUF plugs also showed no significant species indicating that the acetoi
extractions were capable of recovering the analyte species without causing sampl
contamination.
Analysis of the filter and PUF extracts showed that, as expected, the mon
volatile species were found primarily in the PUF extracts. In general, specie
eluting before 4-ethylguaiacol (Figure 1) were primarily in the PUF extract
(although they were also detected in the filter extracts), while later elutiij
species were found primarily in the filter extracts. Table I shows the tota.
concentrations (filter plus PUF extract) of the twelve major methoxylated phenol-
for the three pine samples (A,B, and C) and for the three hardwood samples (D,I
and F). Even though the samples were collected from six different sites, tk
concentration of the individual compounds is surprisingly consistent, particularl;
for the pine samples. While the hardwood smoke had high concentrations of bot
guaiacol and syringol derivatives, the pine smoke had very low levels of syringe!
derivatives. The total concentration of guaiacol derivatives for pine averaged 1
fig/rag, particulate carbon which is in good agreement with the average total guaiaco.
value of 109 ^ug/mg from the hardwood smoke.
Conclusions
Methoxylated phenols have several potential advantages as tracers of atmospherii.
wood smoke pollution in that they should be unique to wood smoke pollution, the;
are present in high concentrations, and they can easily be measured using capillar
gas chromatography/mass spectrometry (GC/MS) without the need for intermediati
class-fractionation steps. Smoke collected from residential wood stoves showe
similar concentrations of guaiacol derivatives whether pine or hardwoods were beiuj
burned, while only hardwood smoke had significant concentrations of syringi
derivatives. These preliminary results indicate that guaiacol derivatives could k
used for tracers of atmospheric wood smoke pollution for both pine and hardwoot
burning, while the type of wood burned could be determined by observing tic
concentrations of the syringol derivatives.
Acknowledgements
The financial support of the U.S. Environmental Protection Agency, Office oi
Exploratory Research (grant number R-813257-01-0) is gratefully acknowledged. EK
also acknowledges the support of the National Science Foundation (NSF ATM-8618793),
References
1. J.A.- Peters, "POM emissions from residential wood burning: An environmental
assessment," Proc.-1981 Int. Conf. Resid. Solid Fuels; Environ. Impacts ant
Solutions, J.A. Cooper and D. Malek, eds. 267-284 (1981).
2. M.V. Buchanan, "Mass spectral characterization of oxygen-containing aromatic!
with methanol chemical ionization," Anal. Chem. 56_: 546-549 (1984).
60
-------
OJ
Table I
Identities and Concentrations of Candidate Tracers for Atmospheric Hood Smoke Pollution
jig per mg particulate carbon**
Species Ret.
Indice3
guaiacol 1093.1
4-methylguaiacol 1196.7
4-ethy Iguaiacol 128 3 . 8
propylguaiacol 1373 . 3
trans-4-propenylguaiacol 1457 . 7
acetonylguaiacol 1540.6
syringol 1356.5
methylsyringol 1452 . 5
ethylsyr ingol 1533 . 0
propylsyringol 1615.5
propenylsyringol 1708.1
acetonylsyr ingol 1781.3
Total guaiacol derivatives
Total syringol derivatives
Mol.
Wt.
124
138
152
166
164
180
154
168
182
196
194
210
Mol. Identity
Formula0 Confirmed0
C7H802 X
C8H1002 X
C9H1202 X
C10H1402
C10H1202 X
C10H1203
C8H1003 X
C9H1203
C10H1403
C11H1603
C11H1403
C11H1404
Pine
ABC
38 41 35
38 38 41
8 14 13
112
234
246
<1 <1 1
<1 <1 <1
<1 <1 <1
<1 <1 <1
<1 <1 <1
89 101 101
<1 <1 1
Hardwood
D E F
72
39
20
4
16
6
3
58
39
13
25
8
157
146
34
21
9
<1
13
8
2
27
38
10
17
7
85
101
52
10
6
2
9
7
1
64
30
10
22
7
86
134
aRetention indices were based on normal alkanes.
^Molecular formulas were determined using high-resolution mass spectrometry.
cldentifications were confirmed by comparing the retention indices and mass spectra with those of known
standards.
°The micrograms of each compound per weight of particulate carbon was determined using GC/MS based on
experimentally measured relative response factors when standards were available. Relative response factors for
the remaining compounds were^estimated based on those determined for the available standards.
-------
o>
Retention Time (min)
Figure 1 GC/FID chromatogram of an unfractionated extract of hardwood soot. The individual components
were identified using GC/MS, GC/HRMS, and deuterated reagent chemical ionization mass spectrometry as
desoribed in the text. Iniection cemper-atuirc? was 8O °C followed by a cempex-ature ramp, to 32O °C jit
-------
OR6ANXC3 DEPOSITION MONITORING:
THE ONTARIO EXPERIENCE
D.B. Orr, P.J. Steer, W.H. Chan, N.W. Reid
Ontario Ministry of the Environment
Air Resources Branch
880 Bay Street, 4th Floor
Toronto, Ontario M5S 1Z8
S.M. Burns and J.Osborne
Ontario Ministry of the Environment
Laboratory Services Branch
125 Resources Road
Rexdale, Ontario M9W 5L1
Abstract
The development of a precipitation and air sampling network
designed to quantify the long-term deposition of organic
compounds into the Great Lakes basin is described. The sampling
methodologies developed have been validated with laboratory and
field quality assurance activities. HCB, a-BHC and g-BHC have
been observed at detectable concentrations in air samples taken
at Dorset, Ontario. Other toxic compounds, such as PCS (total)
and toxaphene have not been detected. Reproducibility estimates
for HCB, a-BHC and g-BHC are good. Target compounds have been
detected intermittently in precipitation samples taken at
Dorset and Port Stanley, Ontario at concentrations near the
minimum measureable amount. Areas for improving the monitoring
effort have been identified.
Introduction
Atmospheric deposition has been identified as an important,
if not the dominant, pathway for the input of anthropogenic
organic compounds into the Great Lakes basin1'5. However,
estimates of organic deposition into the Great Lakes have been
63
-------
based on limited data. For instance, monitoring is often a
short-term, seasonal effort and there is a distinct lack of
airborne organics concentration and deposition data. The Air
Resources Branch of the Ontario Ministry of the Environment
therefore initiated a precipitation and air sampling network
designed to quantify long-term inputs of semi-volatile organics
(substances having vapour pressures roughly between 10"1 and
10~7 mm Hg at ambient temperatures) to each of the Great Lakes
bordering Ontario and to an inland, long-range transport
receptor site. Target compounds include polychlorinated
biphenyls (total PCB), hexachlorobenzene (HCB),
hexachlorocyclohexanes (a-BHC, b-BHC, g-BHC), DDT residues
(o,p'-DDT, p,p'-DDT, p,p'-DDD, p,p'-DDE), chlordanes
(a-chlordane, g-chlordane) , oxychlordane, toxaphene,
heptachlor, mirex, dieldrin, endrin and aldrin.
This paper highlights the sampling and analytical
methodologies developed for the collection and analysis of the
target compounds in air and precipitation and associated
quality assurance activities and presents preliminary air
concentration results obtained by the developed methodology.
Ongoing work to improve the monitoring effort is also
described.
Experimental
Since January 1986, cumulative (28-day) precipitation
samples have been collected sequentially at Port Stanley (Lat.
45°13'26", Long. 78°55'52") on the shore of Lake Erie and at
Dorset (Lat. 45°13'26n, Long. 78°55'52n), an inland long-range
transport receptor site in central Ontario. Additional
precipitation sampling sites, in the proximity of Lake Ontario,
Lake Huron and Lake Superior, were added to the network in
July, 1987. Four-day ambient air samples have been collected
intermittently since January 1986 at the Dorset site only, but
eventually air samples will be added to all of the
precipitation monitoring sites.
A modified version of the automatic wet-only sampler
described by Strachan and Huneault^, as shown in Figure 1, was
used to collect the precipitation samples. With this sampling
train, particle-bound organics were isolated on a 47 mm
soxhlet-treated glass-fiber filter (Gelman A/E) and soluble
organics were trapped on a 70 mm bed of Amberlite XAD-2 resin,
contained within a 15 mm diameter x 300 mm borosilicate
chromatographic column. The effluent passing through the column
was collected for chemical analysis to determine if
breakthrough had occurred. The gravitational flow of
precipitation was regulated at about 40 ml/min.
A modified version of a Sierra-Andersen Hi-Vol air sampler,
outfitted with a Rotron brushless motor, was used to collect
air samples. Air was drawn at 0.35-0.40 m3/min through a 20 x
25 cm Teflon-coated glass-fiber filter (Pallflex T60A 20) which
64
-------
was upstream of a 75 mm diameter x 40 mm stainless steel
cartridge containing approximately 45 g of XAD-2 resin. A
secondary downstream cartridge was used on a trial basis to
determine if breakthrough had occurred. The air sampler
configuration is shown in Figure 2.
The standard operating procedures for both samplers are
described in detail by Steer7'8.
Air and precipitation cartridges and filters were received
at the laboratory within ten days of collection. Upon receipt,
exposed air filters and cartridges were separately soxhlet
extracted in dichloromethane for 12 and 24 hours, respectively.
Precipitation filters were sonified in acetone; distilled water
was added to the acetone and then the mixture was back
extracted in dichloromethane. The precipitation cartridges were
first eluted with acetone, the eluent was combined with
distilled water then back extracted with dichloromethane. The
precipitation cartridge was eluted a second time with
dichloromethane, then both dichloromethane fractions were
combined. Precipitation effluent was extracted with
dichloromethane. All sample extracts were dried by contact with
sodium sulphate and rotary evaporated to 2 ml in the presence
of isooctane. Extracts were then fractionated by florisil
column chromatography into three fractions to separate the PCBs
from the organochlorines .
A Vari'an 6000 dual capillary column, dual electron capture
detector gas chromatograph was used to analyse the extract
fractions. The analysis conditions were: automated splitless
injection; SPD-35 and SPD-5, 60 m x 0.25 mm i.d., 0.25 urn film
thickness capillary columns (Supelco Canada Ltd.); helium
carrier gas at a pressure of 2.6 kg/cm ,• oven program: inject
at 80°C, hold 2 min, program to 270°C at 4°C/min, hold 5 min;
injector 250°C; detector 350°C.
Identification and quantitation was based on external
standards of the individual organochlorines and a mixture of
PCB Aroclor 1254/1260 (4:1). Compound identification required
the presence of the compound on both analytical columns.
Quantitative results were based on peak areas for the single
component organochlorines and the sum of peak areas for multi-
component organochlorines.
Quality assurance (QA) activities are an integral part of
the air and precipitaion sampling program. The following QA
measures were implemented to validate the respective sampling
methodologies :
i) Static field and laboratory recovery tests were done
to assess the stability of PCB, DDT residues and
toxaphene spiked onto air and precipitation
cartridges;
65
-------
ii) A dynamic field test was carried out to quantify
percentage recovery of FOB, DDT residues,
toxaphene spiked onto a precipitation resin column
to assess if breakthrough into the effluent
occurred. For the air sampler, breakthrough
assessed in a similar manner, however, no percent
recoveries were determined since the spike was
radio-labelled.
iii) Blank air and precipitation cartridges and fil
field blanks were a routine feature of each sampl
period to assess handling and transportation effect
iv) By collocating precipitation and air samplers at
Dorset site, overall reproducibility (precis!
estimates for organic compound concentrations W
derived.
Results and Discussion
As indicated in Table I, the percentage recoveries
field and laboratory spikes are quite comparable, suggest
only analytical spikes are necessary in routine netw
operation. Spike recoveries for air samples are better t
those of precipitation samples. The former are typically 90*
greater and the latter are typically around 70%. Recove*
less than 100% are attributed to adsorption onto the sampl
train, especially in the case of precipitation sampling
Franz, gt.Tra4^T9 • observed. The spike concentration levels «
were greater than expected ambient concentrations and it
anticipated that the recoveries observed would be some**
lower for ambient samples.
Breakthrough was not evident neither for air cartridges
precipitation columns spiked with the target compounds, nor
routine samples. This indicates that the samplers effecti?
trap the compounds of interest and thus are conside
acceptable for routine monitoring.
Target compounds on blank filters and cartridges for
and precipitation were invariably non-detectable. It
concluded that transportation and handling protocols do
introduce any contamination problems.
Concentrations of the target organics in precipitat
samples collected at Dorset and Port Stanley were consisted
at or near the minimum measureable amounts with only a
isolated exceptions. The minimum measureable amounts (n?/
based on a ten litre sample, are, for example: PCB (tot*
2.0; HCB, 0.1; a-BHC, b-BHC, g-BHC, 0.1; DDT residues, 0
chlordanes, 0.2; mirex, 0.5; and toxaphene, 20.0. These rest
suggest that the sensitivity of the. analytical method was
adequate for quantifying the trace concentrations of the
compounds and further improvements are needed.
66
-------
Detectable airborne concentrations of a-BHC, g-BHC and HCB
were observed at Dorset (Table II) . Other compounds from the
target list were seen infrequently, only on the resin
cartridge/ and at levels at or near the minimum measureable
amounts. Based on a 2000 m3 sample, minimum measureable amounts
(ng/m3) are, typically: PCB (total), 0.01; HCB, 0 . 0005 ; a-BHC,
b-BHC, g-BHC, 0.0005; DDT residues, 0.0025; chlordanes, 0.001;
mirex, 0.0025; and toxaphene, 0.10. When the filter and
cartridge concentrations of a-BHC, g-BHC and HCB are compared,
it seems that these organics exist predominantly in the vapour
phase. However, it should be noted that, because of sampling
difficulties such as filter collection efficiency and potential
volatilization, the partition of these compounds in to the
particulate and vapour phases might not be representative. The
results reported in Table II are similar to measurements made
in Europe and North America as reported by Bidleman, et al.10.
(and references therein). The reproducibility of the organics-
in-air results is quite good, ranging from 72% for a-BHC an.d g-
BHC to 97% for HCB, as shown in the last column of Table II.
Ongoing Work
Areas for improving the monitoring effort have been
identified:
i) The analytical sensitivity for precipitation samples
are being improved by at least an order of magnitude
since most of the organochlorines detected were at or
near their minimum measureable amounts. Increased
sensitivity will be accomplished by increasing the
analyte concentration factor. Alternative sample
fractionation processes and the use of internal
standards are being investigated.
ii) The target compound list has been shortened by
removing toxaphene, dieldrin and endrin. Quantitation
of PCB will be based on 77 specific PCB congeners
instead of the total PCB Aroclor approach.
iii) Static laboratory recovery tests and dynamic field
recovery tests for air and precipitation cartridges
and filters spiked with the 77 PCB congeners, HCB,
a-BHC, b-BHC, g-BHC, a-chlordane, g-chlordane,
oxychlordane, heptachlor, aldrin and mirex will be
implemented after the analytical sensitivity has been
improved.
iv) In future, a methanol rinse will be flushed through
the collection funnel and cartridge of the
precipitation collector prior to collecting the; filter
and cartridge samples to minimize the retention of
target compounds on the sampling train.
67
-------
v) Equipment modifications to improve sample collection
efficiency will include: a) a higher capacity pump for
the air sampler regulated by a mass flow controller)
b) an extension below the resin cartridge to stabilize
the passage of air past the flow controller probe; c|
a deeper collection funnel for the precipitation
collector to limit rain splash; d) a Teflon gasket
lining the underside of the moveable hood of tlw
precipitation collector to minimize evaporation and to
limit wind-blown dust contamination and; e) an
extended Teflon funnel adaptor to stabilize the level
of precipitation in the sampling train.
vi) Additional air samplers are to be collocated at the
existing precipitation collection sites. This will
improve the understanding of the spatial variability
of organics deposition within the Great Lakes basin
and will provide information on scavenging ratios.
References
1. • W.M.J. Strachan, "Organic substances in the rainfall
of Lake Superior: 1983," Envir. Sci. Technol. 4_: 671
(1985) .
2. T.J. Murphy, Toxic Contaminants in tjie Great Lakes. J,
Wiley and Sons, New York. 1984, pp.53-79.
3. W.M.J. Strachan, S.J. Eisenreich, "Mass balancing of
toxic chemicals in the Great Lakes: the role of
atmospheric deposition," International Joint
Commission workshop on atmospheric loadings of toxic
chemicals to the Great Lakes basin, Scarborough,
Ontario, (1986) .
4. W.M.J. Strachan, H. Huneault, "Polychlorinated
biphenyls and organochlorine pesticides in Great Lakes
precipitation," J. Great Lakes Res__,_ £.: 61 (1979) .
5. S.J. Eisenreich, B.B. Looney, J.D. Thornton, "Airborne
organic contaminants in the Great Lakes ecosystem,"
Envir. Sci. Technol . 15 : 30 (1981) .
6. W.M.J. Strachan, H. Huneault, "Automated rain sampler
for trace organic substances," Envir. Sci. Technol,
18 : 127 (1984) .
7. P.J. Steer, "Standard operating procedures for the
sampling of organics in air. Internal technical
memorandum, Air Resources Branch, pp.6, (1986a).
68
-------
P.J. Steer, "Revision to standard operating procedures
for the sampling of organics in precipitation.
Internal technical memorandum, Air Resources Branch,
pp.10, (1986b).
T.P. Franz, M.B. Swanson, S.J. Eisenreich, "Field
intercomparison of rain samplers for assessing wet
deposition of organic contaminants," International
Association for Great Lakes Research 31st Conference,
M 1Afl S
(1988) .
11
T.P. Bidleman, U. wideqvist, B. Jansson, R. Soderlund,
"Organochlorine pesticides and polychlorinated
biphenyls in the atmosphere of southern Sweden,"
Atmoa. Envir. 21: 641 (1987) .
A.J.S. Tang, W.H. Chan, D.B. Orr, W.S. Bardswick, M.A.
Lusis, "An evaluation of the precision, and various
sources of error, in daily and cumulative
precipitation chemistry sampling," Water. Air. and
Soil Poll. 36: 91 (1987) ,
69
-------
Table I Laboratory and field recovery testa for spiked cartridges.
- Precipitation --
Amount
Air
Amount
Compound
PCB (total)
p,p'-DDE
0,p'-DDT
p,p' -ODD
p,p' -DDT
Toxaphene
Type of
Spike
Lab static
Field atatlc
Field dynamic
Lab static
field static
rield dynamic
tab static
Field static
Field dynamic
Lab static
Field static
Field dynamic
Lab static
Field static
Field dynamic
Lab static
Field atatlc
Field dynamic
Spiked1
Minimum (I)
a-BHC
g-BHC
HCB
filter
Cartridge
Filter
Cartridge
Filter
Cartridge
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.198
<0.001
0.025
<0.001
0.086
0.002
0.534
0.005
0.127
0.005
0.150
9.4
86.7
6.3
73.3
2.9
73.3
ND
72.3
ND
72.2
NO
96.8
ND Not determined; insufficient paired data > minimum.
1 Results not corrected for percent recovery.
2 n - 30.
3 Reproduclbillty is as defined by Tang,
used.
15 matching pairs
70
-------
TOXIC CHEMICALS IN CANADIAN RAINFALL
William M. J. Strachan
National Water Research Institute,
Canada Centre for Inland Waters,
P. 0. Box 5050/867 Lakeshore Road,
Burlington, Ontario, Canada, L7R 4A6.
Polychlorinated biphenyls and organochlorine pesticides have been observe
in the rain and snow in Canada since before 1980. Recent efforts hav
expanded the list of compounds investigated to include the chlorobenzene
and a selected number of polynuclear aromatic hydrocarbons. Th
concentrations of all of these chemicals are determined in rainfall usij
large surface area (0.2 nr) wet-only collectors and resin sorption colum
Triplicate samples of rainfall have been collected from a total of te
sites ranging from British Columbia on the west coast to New Brunswick o
the east. Generally, the bulk of the rainfall period was sampled at eac
location.
Only the data for PCBs and CCs have sufficient quality to discuss at thi
time; other compounds are observed but the levels and interferences ar
such that only qualitative statements can be made about their presenci
For the PCBs and OCs, however, a similar pattern is found at all site!
The hexachlorocyclohexanes (lindane and the alpha-isomer) are the moa
prominent and PCBs are always present, albeit at lower levels; dieldrii
DDE and hexachlorobenzene are also frequently observed. Concentratic
levels and loadings are presented and the fluxes of PCBs in the Great Late
discussed in the context of the significance of the atmospheric rout
relative to other input mechanisms.
KEYWORDS: Rain, toxic, atmosphere, Canada, PCBs, pesticides.
72
-------
INTRODUCTION
That man-made persistent chemical substances are present in the aquatic
environment and elsewhere is not a question; what has become of concern is
the role that the atmosphere plays in putting them there. There are many
reports, some dating back to the mid-sixties, of observations in
environmental samples which are directly or highly suggestive of the
involvement of atmospheric mechanisms. In Canada, concern over the
deposition of a number of persistent organic chemicals dates from the mid-
seventies when polychlorinated biphenyls (PCBs) and some organochlorine
pesticides (OCs) were observed in rainfall in the Great Lakes area [1-3].
Since then, a number of reports have appeared on levels of these chemicals
in the rain, the air and somewhat in accumulated snowfall [4].
Present studies have had a two-pronged, approach — a statistical one where
triplicate samplers are sited at selected locations for one or two years,
and, a monitoring one where individual samplers are sited on a continuing
basis. It is the replicate programme which is the subject of this report
and its intent is to gather data to convince the monitoring agencies to
undertake and support the development of the nonitoring network. It is also
intended to encourage other investigators to ensure that the present,
binational efforts in the Great Lakes area are expanded and that the
problems defined are addressed on a multinational basis.
EXPERIME3STTAL
The samplers and analytical methodology have been described in a paper by
Strachan and Huneault [5] Rainfall is collected by a 0.2 m 2 teflon-coated
funnel with a automated lid to capture wetfall only. The rain is passed
through a teflon column of XAD-2 resin and the analytes are sorbed and
subsequently evaluated using gas chromatography. In the work on rainfall
in Canada reported here, all results noted were derived from volume
weighted means of a number of triplicate samples covering the bulk of the
rainfall period (generally May - October/November). The locations are
shown on the map Figure 1.
The analytes selected prior to 1986 were the PCBs and approximately 15 OCs.
The sensitivity of the methodology was generally of the order of 0.02 ng/L
for the pesticides and about ten times that for the "class compound" PCBs.
In 1986, additional compounds were added to the analyte list — selected
from the chlorobenzenes (CBs) and polynuclear arometic hydrocarbons (PAHs).
tfo quantitative results are presented for these compounds, however, since
difficulties with blanks have made the quantitation to date dubious.
RESULTS
Concentrations in the rain are presented in Table 1. The data presented
are for the most prominent chemicals observed from among those investigated
and occur in most samples regardless of location or season. There were
other compounds observed but not so frequently and they are omitted here.
Among these other compounds, most were at levels less than 1 ng/L and
included heptachlor epoxide, the two endosulfan isomers, endrin, DDT
itself, and methoxychlor. The later pesticide was found frequently in Lake
Superior samples at concentrations frequently exceeding 1 ng/L but was
73
-------
seldom observed in samples from other locations.
In addition to the PCBs and OCs mentioned, a number of CBs and PAHs wen
observed. While these are proving difficult to quantitate at environmental
levels, a number of them can be qualitatively said to be present at levels
of 1 ng/L or greater including: o- and p-dichlorobenzenes, 1,2 3- and
1,2,4-trichlorobenzenes, 1,2,3,4-tetrachlorobenzene and pentachlorobenzent
observed frequently and 1,2,3,4-tetrahydronaphthalene, phenanthrene and the
methylnaphthalenes which are found less commonly.
The concentration pattern for the compounds in Table 1 is very similar fa
all of the sites. The hexachlorocyclohexanes, the alpha-isomer and
lindane, occur at the highest concentrations and average 16. ng/L for the
alpha-HCHand 4.7 ng/L for lindane (the gamma-HCH isomer) for all sites.
PCBs, which are much noted contaminants in the aquatic part of the
ecosystem, are also prominent in the rain with a Canada-wide mean of 3.{
ng/L. Other frequently observed compounds include the pesticide dieldrin
(also derived from the organochlorine pesticide, aldrin), p,p'-DDE (one of
the degradation products from DDT) and hexachlorbenzene (HCB) which is both
a pesticide and a by-product in organochlorine solvent production.
A preferred way to examine the data is to deal with loadings rather than
concentrations. It is easy to perceive that concentration differences my
be due to dilution at particular locales or during years with different
precipitation intensities. Consequently, data for concentrations are
presented in Table 2 as loading rates with those for the two Superior
locations in 1983 and 1984 being averaged. A similar pattern evolves as
exists for the concentrations except that the west coast appears to be the
recipient of a higher loading of the HCHs per year than is the case for
other parts of Canada. This may be a consequence of high usage of
lindane/HCH in the far east but this is unsubstantiated. There are
however, reports of atmospheric deposition of this set of chemicals in
Japan and Hawaii. It may also be the result of the higher precipitation in
this region with a rapid resupply and subsequent washout fron equivalent
air masses.
The atmospheric deposition of all of the chemicals in Tables 1 and 2 was
the subject of a recent workshop on the subject [6]. This was sponsored by
the Internatinal Joint Commission, a body which has the responsibility of
settling boundary water disputes and administering the Canada-U.S. Great
Lakes Water Quality Agreemnet. A particular question addressed was the
relative significance of the atmospheric route for any or all of twelve
substances in any or all of the five Great Lakes. Among the organic
compounds, only the data base for the PCBs was adequate to address all
lakes and even then there were data gaps,
A simple mass balance accounting (Figure 2) was used with the l)est -workshop
judgement on concentrations and relevant parameters. It is apparent that
PCBs are arriving mainly from the atmosphere; Table 3 summarizes the
sources. The reduced significance for the lower lakes was not so much due
to a lowered deposition there but to the fact that there were substantial
inputs into the connecting channels (Detroit/Lake St. Clair/St. Clair river
system for Erie and the Niagara river for Ontario). These inputs are
believed to be mainly industrial and landfill seepage although there maybe
some input via feeder streams to these river systems.
74
-------
AND CONCLUSIONS
It is apparent that persistent organochlorine chemicals are being deposited
to surface waters and presumably elsewhere via the rain. There is some
evidence on PCBs in the air (mainly in the vapour state) of the Great Lakes
region and it is probable that this is true also for the other compounds.
A major question arises as to the significance of this presence. Prom the
exposure perspective in the aquatic environment, the atmosphere appears to
represent the major source of a number of compounds which bioaccumulate to
levels which result in the need to control human intake of the fish
involved. This is true in the Great Lakes but may also be the case
elsewhere. Even for PCBs and the few other substances, however, the data
do not exist to actually substantiate this claim — they are only adequate
to indicate that this is probably the case.
Despite nearly two decades of intensive monitoring, the concentrations over
sufficient periods of time to determine trends are available only for
levels in biota; the basic compartments of air (rain, snow, air-borne
particulates and vapour) and water (dissolved, sorbed to suspended matter
and sediments) have not been investigated sufficiently. This is due, in no
small measure, to our previous inability to analyse on a routine basis in
these sub-compartments. This is no longer the case and studies should now
be undertaken to provide a reliable baseline on which to base comparisons
and establish trends for raeasuriag thn success of control measures.
REFERENCES
1. W. R. Swain, "Chlorinated residues in fish, water and precipitation
from the vicinity of Isle Royale, Lake Superior," J. Great Lakes Res.
4: 398 (1978).
2. T. J. Murphy, C.P.Rzeszutko, "Precipitation inputs of PCBs to Lake
Michigan," J. Great Lakes Res. 3: 305 (1977).
3. W. M. J. Strachan, H. Huneault, "Polychlorinated biphenyls and
organochlorine pesticides in Great Lakes precipitation," J. Great
Lakes Res. 5: 61 (1979).
4. S. J. Eisenreich, B. B. Looney, J. D. Thornton, "Airborne organic
contaminants in the Great Lakes ecosystem," Envir. Sci. & Tech. 15:
30 (1981).
5. W. M. J. Strachan, H. Huneault, "Automated rain sampler for trace
organic substances," Envir. Sci. & Tech. 18: 127 (1984).
6. W. M. J. Strachan, S. J. Eisenreich, "Mass balancing of toxic chemicals
in the Great Lakes: The role of atmospheric deposition," Appendix 1
from the Report of the Workshop on the Estimation of Atmospheric
Loadings of Toxic Chemicals to the Great Lakes Basin, International
Joint Commission, Windsor, 157 pages (in press).
7. W. M. J. Strachan, H. Huneault, W. M. Schertzer and F. C. Elder,
"Organochlorines in precipitation in the Great Lakes Region," in
Hydrocarbons and Halogenated Hydrocarbons in the Aquatic Environment
(1980), ed. B. K. Afghan and D. Mackay, Plenum Publ., p.387 (1980).
75
-------
Table 1: Mean Concentrations of Contaminants in Canadian Rain
Lake Superior - - - - -
Compound I.Jtoyale Caribou Caribou Agawa
1983 1983 1984 1984
N.B. N. Sask.- - - Alta. B. C. -
Kbuch. Cree L. Cree L. Suffield Kanaka KanaJ*
1984 1984 1985 1985 1985 1906
a-BHC
Lindane
Dieldrin
P,P'-DDE
PCB's
HCB
36.
8.4
nd
0.37
6.7
0.03
15.
4.3
0.24
0.08
5.9
0.10
6.5
3.0
0.96
0.17
2.5
0.09
concentrations in nanograms per Litre
6.5
3.0
0.96
0.17
2.5
0.09
6.7
2.9
0.62
0.09
3.2
0.03
13.
6.7
0.27
0.02
1.1
0.07
6.5
1.2
0.38
0.07
3.1
0.01
22.
6.5
0.04
0.04
3.5
0.01
14.
5.9
0.10
0.03
5.5
0.84
29.
5.0
0.27
— —
_ _
- -
Table 2s_ Contaminant Loading Rates from Rainfall
Rain (mm)
Snow (irni)
—Lake Superior—
1981r 1983 1984
543
164
607
214
618
263
Kbuch.
1984
1050
377
-Cree Lake-
1984 1985
307
176
274
207
loadings in micrograms/m^/annum
a-HCH
Lindane
Dieldrin
p,p'-DDE
FCBs
HCB
9.5
3.3
0.15
(0.08)
1.2
11.
3.7
0.36
0.06
4.9
0.05
3.9
2.0
0.50
0.08
2.6
0.04
14.
7.3
0.30
0.02
1.6
0.08
2.1
0.39
0.12
0.02
1.5
tr
6.5
1.9
0.012
0.012
1.7
0.003
Suffield
1985
208
98
3.3
1.3
0.022
0.007
1.2
0.18
- Kanaka Cr.-
1985
1241
56
36.
6.2
0.34
— _•
- -
- -
1986
1651
34
22.
4.8
0.03
0.03
1.0
0.05
. - replicate data as others except only covers 20 % of wetfall period.
* - snownelt concentrations 100% rain for FCBs and 10% for pesticides [7].
Table 3: Sources of PCBs to the Great Lakes
Direct Atmospheric Connecting Channel Other Direct
Dryfall Wetfall Atmosphere Other Tributaries Discharge
Percent Contribution
Lake Superior
Lake Michigan
Lake Huron
Lake Erie
Lake Ontario
39
25
27
3
2
52
33
36
4
3
15
6
1
7
69
82
9
42
8
18
12
LT 1
?
7
76
-------
Figure 1: Sites in Canadian Replicate Rain Sampling Programre
Figure 2: Framework for Fluxes of Chemicals in an JSquatic Ecosystem
Rainfall
Dryfall
Gas Exchange
R.T-
Fr=Cr.P.SA Fd=Ca.(l-fv).Vd.SA Fv=K.(Cw-[Ca.fv.!ill]).SA
Tributary
Connecting
Channel
I
1
Fs " css'waccifslSA
j
Sediments
Outflow
Data
C's
f£ -
H -
K -
P -
Q's -
R -
SA -
T -
W
acc
Requirements:
concentrations in rain, air (vapour plus particulate) ,
water (dissolved), suspended solids, tributaries
(total) and connecting channels (total)
fraction of atmospheric contaminant present as vapour
fraction of lake area with significant deposition
Henry's Law constant
bulk (or net) mass transfer coefficient
precipitation to lake surface
flows in tributaries and connecting channels
gas constant
surface area of the lake
surface air temperature
particulate deposition velocity
average sediment accumulation rate (deposition zones)
77
-------
AN UPDATE ON GRAB SAMPLING OF VOLATILE ORGANICS
(VOC'S) AND OTHER TOXIC GASES
Joseph P. Krasnec
Atmospheric Chemistry Research Group
Scientific Instrumentation Specialists
Abstract
This paper intends to provide an overview and to describ
recent developments in the grab sampling of Volatile Organii
Compounds (VOC's}. Wide acceptance and successful application o:
this technique is underscored by the inclusion of this approacl
in the U.S. EPA's Compendium (EPA-600/4-84-041) as an approve:
sampling and analysis procedure. This new Compendium Method (TO
14): Determination of Volatile Organic Compounds (VOC's) i:
Ambient Air Using Summa Polished Canister Sampling and Gas
Chromatographic Analysis (2) provides the environmenta!
monitoring and regulatory agencies as well as other interests
users with specific guidance for the sampling and measurement of
selected toxic organic compounds in ambient air. The purpose o:
this paper is to provide additional technical information on th'
methodology, sampling instrumentation and other relate;
technologies. Manual and automated, sequential grab sampling
systems, sampling canister recycling systems, and relate!
hardware is described and discussed in some detail. In addition
sampling instrumentation testing, certification and validatiot
procedures, and applicable analytical methods are presented.
Introduction
Recent developments in the design of rigid grab samplinf
containers led to an increased use of sampling devices made fro:
specialised materials, i.e. passivated stainless steel (PSS
resulting in successful sampling and storage of trace and toxic
gases at the low part-per-billion (ppb) and part-per-trillioi
(PPt) level for periods of time ranging from a few days tc
several months or longer (4). Advances in the construction of PSS
grab sampling containers coupled with improvements in the
sampling procedures and availability of automated, modular gral
sampling systems make this approach to toxic air pollutant
sampling increasingly attractive. Information is now available
on the stability and storability of up to two hundred organic
compounds as research data from several studies is presented and
published during 1988. Several large, on-going studies have
successfully used grab sampling with PSS sampling container:
confirming viability of this sampling method.
Experimental Methods:
Grab Sampling Instrumentation Design & Construction Requirements
Most important requirements for grab sampling containers,
78
-------
samplers and accessory hardware are a non-
non-reactive (inert) contact surface; minimal
to volume ratio; reusability and durability; suitable
p . anc* construction; acceptable size and weight
°rtabilityj; simple sample collection procedures; acceptable
. > safety and ease of use; and application to a wide variety
gaseous pollutants.
§r Proper, extensively tested and certified construction and
:rit' P"L°yment handling of pss grab sampling containers is
^ ical because these parameters determine the degree of success
US*ng this technique for VOC sampling and analysis. While
vely simple, the PSS grab sampling containers evolved into
icated, commercially produced sampling instruments. The
aPPr°ach, selection of high quality components and
, t QG/QA assembly and testing/certification procedures are
^^ itraost importance. The passivation os one of the key steps in
^ manufacture of grab sampling containers. It is much more
•eth than electropolishing or other surface cleaning
Uj-f- S' SUMMA passivation removes surface scale, decreases the
O ace contact area, removes impurities and active sites, and
an inert nickel/chromium oxide layer on the stainless steel
surface. The cleanliness and inertness of SUMMA
etted SS sampling containers has now been well documented
as nuniber °^ short and longer term stability/storability
es have been completed and data has been reported.
ont . t^ proper cleaning (recycling) of the PSS sampling
•*te t*ers the QA/QC analytical blanks are below the limits of
10n' for GC and GC~MS instruments. Certification of the grab
containers is thus not only possible but it needs to
an essential requirement for all grab sampling and
tasks- Specifically, each PSS grab sampling container,
automated sampler and other accessory hardware should
vigorous cleaning and certification procedure to assure
Blanks and valid data. An unique I.D. number for each
e coupled with a permanent use log, (re)certif ication and
Us^/sampling parameters is highly recommended.
aropling Instrumentation Cleaning and Recycling
used grab sampling container cleaning/recycling
mjs es utilise evacuation to about 5 mm Hg (5 torr) and
hig quen"t pressurisation with humid zero air to 30 psig (2).
:VQC Procedure is repeated two more times for a total of three
Of ..5 *on/pressurization cycles. The blank(s) are then obtained
JJe desired containers by FID-ECD/GC analysis. Typically, oil
vane pumps are used for pump-down to as low as 10-3
pressure measurement is to about 5 torr. The pump
for lower pressure is limited by the vacuum gauge
mechanical or hot wire type. It is highly
to pump-down grab sampling containers to 10-3 torr or
vacuum to remove as much of the remaining sample as
e^ As an illustration, at ambient pressure there are 2.5 x
in one cm3, about 3.3 x 10+8 monolayers/sec and a
79
-------
mean free path of 6,3 x 10-5 cm. At 10-3 torr there are 3.5
10+13 molecules or one million times fewer molecules per cm3 , 4<
monolayers/sec and a mean free path of 5 cm which in both cas(
is again about a million-fold (10 + 6) decrease. This , of couis
is very desirable for the purpose of thorough, reliable cleani:
of all grab sampling hardware. Considering the ppb or ppt I
concentrations to be measured in collected air samples, a
possible contamination from sources at a ppm or high
concentration the desired million-fold or higher dilution duri:
the pump-out is essential.
In some cases heating the containers to temperatures up
about 400 F may be required to remove more reactive organ!
that are adsorbed on the SS surface. The adsorption typical
'takes place because of the physico/chemical characteristics of
given compound, such as molecular weight, boiling point, vap
pressure, presence of C=C or C-0, 0-H or other chemical bonds a
properties. Commercially available laboratory ovens can
modified to accomodate up to eight containers attached to
common manifold. Typical bake-out time ranges from 6-16 hours
temperatures between 200 and 400 " F. Concurently., the enti
system needs to be evacuated to at least 10-3 torr with
adsorbant or LN2 trap between the pump and the pumping manifold
Grab Sampling Instrumentation Certification
The grab sampling container certification requires reliabl
accurate and traceable analysis (blanking) following the clean!
procedure. Ideally, the analysis should be performed during
before .completion of the cleaning procedure. This can be da
with a quadrupole mass spectrometer Residual Gas Analyzer (fiGJ
directly connected to the cleaning system. Such system :
currently used in a commercial application. The major componen:
are a high speed roughing pump section for system pump-down '
10-3 torr, secondary pumping system utilising a diffusion pu:
for pump-down to 10-5 torr5 and a final stage with a molecul:
drag pxunp for ultimate pressure of 10-6 torr. Both foreline ai
LN2 traps are used for isolation of the pumps and the sampli:
containers. The EGA analyzer head is installed in the high vacui
section of the system (10-4 to 10-6 torr) for optiit,
performance. Computer conti-ol, real time display of desirt
paramaters and ability to obtain a hard copy for report!:!
purposes are some of the advantages of this system. The S'-l
allows simultaneous monitoring uf up to 16 different gases ai[
provides a continuously updated status report on their pressure
(concentration) in the cleaning system. In addition, it can Ij
used in a helium leak detection mode. Follow-up GC or GO!';
analysis can be performed, if required, by back-filling the gn|
sampling containers with analyzed zero air, storing the contain?
for at least 12 hours and then withdrawing a sample for analysis
Instrumentation for Manual and Automated Sampling of VOC ' s
Manual grab sampling typically utilizes only a
container of a specified size for essentially an instantaneod
80
-------
collection of air samples. The container is evacuated and
certified prior to use. If time-averaged sample collection is
required a number of flow controlling devices can be utilized.
Fixed flow orifices, needle valves, pneumatic flow controllers,
mass flow controllers and fixed volume metering pumps are
available. Most of these flow control devices require a pressure
differential (1-10 psig) across the unit for proper flow control.
Sub-ambient and ambient pressure sampling is carried out without
a pump. This approach does not require electric power, eliminates
possible contamination or sample carry-over by the sampling pump,
and offers a simple, yet effective means of collecting air
samples in almost any environment. Recent research data seems to
indicate that longer term stability of VOC's is best at ambient
or near ambient pressure. Portable, self-contained and ready to
use manual grab sampling systems are commercially available.
Positive pressure sampling utilizes a clean SS bellows, SS
pump head/Viton diaphragm pump or a novel low flow metering pump.
A desired pressure differential across the flow control device
is easily maintained, and a positive pressure (10-30 psig)
sample is collected. Both sub-ambient/ambient and positive
pressure manual sampling systems can be equipped with
vacuum/pressure gauges, inlet on/off valves and manifolds for
multiple grab samples. Typically, a single valve grab sampling
container is used in sub-ambient and ambient sampling. If
required, an integral vacuum/pressure gauge can be installed
directly on the grab sampling container. Also, a single valve
configuration can be changed to a two valve configuration with a
purge tee assembly, providing that the grab sampling container
uses a modular construction. Positive pressure sample collection
can utilize both single valve containers, and two valve
containers for dynamic sample flow through the sampler. The exit
valve is closed at a predermined point and the sampling container
is pressurized to a desired pressure. The two valve sampling
container configuration provides an additional advantage for
container cleaning by allowing purge flow of zero air or other
clean gas, i.e. clean background air through the container. This
cleaning approach should be used only for cleaning/recycling
samples containing low (background) concentrations of VOC's.
Automated grab sample collection systems use electronic
timers (6) or in some cases microcomputers for automatic control
of the sampling cycle. Typically, a timer is set to start and
stop the sampling sequence by opening/closing valve(s) and
starting/stopping pump. Commercially available automated single
container sampling system utilizes a digital seven day timer with
four user selectable sampling sequences that provide a choice of
start/stop time and day. In addition, a duty cycle of 1-100% can
be selected on the control panel with thumbwheel switches. The
duty cycle uses a fixed time base, i.e. 5 minutes. With 100% duty
cycle setting the sampler pump will operate continuously and the.
electrically operated three-way latching solenoid valve will be
open for the entire sampling period. With a 50% duty, cycle
setting the sampler pump will be turned on and the solenoid valve
switched to an open position for 50% of the 5 minute time base
81
-------
(2.5 minutes) throughout the entire sampling cycle. During tl
remaining 50% of time the sampling system is purged. USE
selectable duty cycle allows collection of samples at higher flo
rates as compared with continuous sampling for the same lengthc
time. Also, smaller, more compact sampling containers can be use
when compared to continuous sampling at the same flow rate
Another advantage of this sampler is its ability to provide sub
ambient/ambient (pumpless) sampling for longer term (i.e. 24 k
or more) sampling cycles. With the duty cycle control feature th
actual sample collection takes place for time intervals as shor
as 3 seconds every five minutes (1% duty cycle setting- 1% of 3t
seconds). The above described model of the automated gra.
sampling system is modular, and it can accomodate samplit
containers in sises from 0.5 1 to 15 liters. The units are full;
portable, with largest module for 15 liter sampling container
measuring 25"xl2"x20" and weighing about 30 pounds. Optionally
they can use a 12V DC power source.
The automated, microcomputer controlled grab samplin
systems also utilise modular construction that allows use of 0.:
1 to 15 liter grab sampling containers. The Sharp 1500A/160
microcomputer provides full control of all sampling parameter;
(individual start/stop time, interval between samples, purgr
time; solenoid valve sequencing; sampling data compillatioi
storage and output [LC screen display, hard copy print-out «:
transfer via optional RS-232C interface], and manual/diagnosti
operation of the system) . Compared to the digital timer operate:
single container sampler the microcomputer controlled model ca
sequentially collect up to sixteen samples. Additionally, a pur£
sequence is programmed into the sampler operating software tha:
provides for user selectable setting of purge time (prior t
actual sample collection) for each sample. Also, a sample flo;
bleed valve is provided for adjusting the pump head pressure. Th>
master unit with the computer, electronics, power supply, pump
flow controller and purge bypass is connected to the stand-aloft
slave units by means of sample flow connecting lines and solenoi:
valve cables. The slave units contain additional grab samplirif
containers, latching solenoid valve for each container
downstream isolation valve and the manifold connecting tht
sampling containers. Both master and slave units are modular
with removable lids, and are fully portable.
Some automated samplers include sample inlet line and ON/OFi
valve, particulate filter, vacuum/pressure gauge(s), samplert
container connecting lines, and separate compartments for eac:
sampling container. 12 VDC operation is possible with an options
power cord (samplers are powered by an internal 12 VDC powe:
supply). Flow control is provided by a high quality^ ta.mpet
proof, stainless steel flow controller with a flow adjustment i:
the 5 to 500 ml/min range. The flow controller can be easily set
and calibrated in the field and will reliably hold set flow fo:
extended lenght of time. The latching stainless steel solenoi:
valves have minimal power consumption because they are energize:
only for fraction of a second to open or close. All valves ar?
helium leak tested. Compact, low power sampling pump has ;
82
-------
stainless steel pump head and a Viton diaphragm. All connecting
tubing, fittings, vacuum/pressure gauges and sampling container
nmif olds are made from stainless steel .
Conclusions
The grab sampling method using Summa passivated stainless
steel containers and GC ; GC-MS analysis (2, 3) has firmly
established itself as the most reliable and preferred method for
measurement of Volatile Organic Compounds { VOC's) in ambient air.
This method has become an approved method (TO- 14) in May 1988 and
is included in U.S. EPA's Compendium of Methods for the
Determination of Toxic Organic Compounds in Ambient Air.
Commercially produced PSS grab sampling containers, manual and
automated samplers, turn-key container cleaning/recycling systems
and associated accessories are now available to support the
ongoing and planned sampling projects. Sampling methodology and
procedures have been established to provide for reliable,
reproducible and traceable sampling of VOC's in ambient air.
QA/QC protocols are being finalized to support and validate
sampling in many different settings, ranging from toxic waste
disposal sites to industrial and indoor atmospheres. Analytical
techniques are available to perform sophisticated multidetector
GC analysis, or sensitive and reliable VOC speciation by MSD/GC .
References
1. K.D. Oliver, J.D, Pleil , and W.A. McClenny, "Sample Integrity
of Trace Level Volatile Organic Compounds in Ambient Air Stored
in Summa Polished Canisters", Atmospheric Environ. 20:1403, 1986.
2. F. McElroy, "Compendium Method TO-14: Determination of
Volatile Organic Compounds (VOC's) in Ambient Air Using SUMMA
Polished Canister Sampling and Gas Chromatographic Analysis",
Draft, U.S. EPA/EMSL , Research Triangle Park, N.C. 27711, 1987.
3. W.T. Winberry and N.V. Tilley, "Supplement to EPA-600/4-84-041 :
Compendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air", EPA-600/4-87-006 , U.S. EPA, Research
Triangle Park, N.C. 27711, 1986.
4, J.P. Krasnec, "Grab Sampling as an Effective Tool in Air
Pollution Monitoring", Proceedings of the 1987 EPA/APCA Symposium
on Measurement of Toxic and Related Air Pollutants, Research
Triangle Park, N.C., 1987.
5. W.A. McClenny, J.D. Pleil, M.W. Holdren, and R.N. Smith,
"Automated Cryogenic Preconcentration and Gas Chromatographic
Determination of Volatile Organic Compounds", Anal. Chem.
56:2947, 1984.
6. W.A. McClenny, J.D. Pleil, T.A. Lumpkin, and K.D. Oliver,
"Update on Canister-Based Samplers for VOC's", Proceedings of the
1987 EPA/APCA Symposium on Measurement of Toxic and Related Air
Pollutants, Research Triangle Park, N.C., 1987.
83
-------
MEASUREMENT AND EVALUATION OF
PERSONAL EXPOSURE TO AEROSOLS.
R. W. Wiener
Environmental Monitoring Systems Laboratory
U, 5. Environmental Protection Agency
MD-56
Research Triangle Park, NC 27711
The Particle Total Exposure Assessment Methodology (Particle-TEAM)
Program of the U.S. Environmental Protection Agency (USEPA) is
concerned with developing the means by which an estimate of the
frequency distribution of human exposure to aerosol particles can be
made for a population. Sampling instruments, questionnaires,
methods, and protocols are being developed for use in Particle-TEAM
field sampling programs. Preliminary studies have been conducted to
evaluate the performance of some of these instruments and to begin to
discern the sources of human particulate exposure.
84
-------
The Particle Total Exposure Assessment Methodology (Particle-TEAM)
program IB currently being developed by the U.S. Environmental
tu°tection Agency (USEPA) to estimate the level of human exposure to
Particles and relate exposure to sources of aerosol matter. This
ev°gr&n f°H°ws a sequence of TEAM programs designed to estimate
eryday exposures to other potential hazards such as volatile organic
conpounds (VOC's). It is designed to utilize information and
r-chnologies obtained and/or developed from a variety of EPA programs
Deluding the Indoor Air Program. Identification of exposure
con* tions and sources is essential in the development of realistic
°"Jfo1 strategies for reducing risk from human exposure to
see 4 Ulate Batter indoors. The Particle-TEAM program is focused on
»et i E*ze fractions of aerosol particles analyzed for at least
^»ls, nicotine, and as resources permit, semi-volatile organics.
numhTEAM Pr°9rans seek to answer fundamental questions regarding the
DOII r °* Persons exposed, the sources of exposure, transport of
tteai^tant to the population at risk, the effects of exposures, the
est< ^ of BaaPling data in terms of actual exposure, and the
boil tion of the l«vel of exposure of the actual population to the
rent in Question. The basic ingredients of TEAM study are
con tative Probability sampling, measurement of the pollutant
Det. entrations, measurement of body burden, and recording of each
anew '* dailv activities.l»2 The Particle-TEAM program seeks to
ext^A r <^uestions relating the measurement of sources to human
vii?8ure* Health related Information collected in this TEAM study
Vill be limited.
d«ve?he first 8t«P ln initiating the Particle-TEAM program is the
ftonii°pttent and selection of sampling equipment for personal exposure
detoei ng CpEM) and/or fixed location microenvironmental sampling of
that ? An a«rosol PEM consists of a flow controlled personal pump
acti ** worn by the test subject as he/she performs his normal daily
and &n inlet and filter pack which are designed to collect
e Particles (PM2.5 or PM10) near the breathing zone. Pump
o .
8tavj?!?"niyersity of Minnesota for the measurement of personal exposure
C°nsi5? blc (pM2-5) or inhalable (PM10) particles.3 Each inlet
ptedef of an imPactor classifier to remove particles larger than the
I>art?^ernined cut size, and a filter to collect the remaining
the i^T * Two disassembled inlets are shown in Figure 1. Four of
*7ttin *T?ts were designed to operate with 37nun filters and four with
niters. The inlets were designed to operate at either 4 LPM or
85
-------
10 LFN flow rate and have cut sizes of 2.5 or 10.0 urn aerodynamic
particle diameter. These cut sizes were chosen to represent the
respirable and inhalable aerosol fractions. Currently the outdoor
ambient standard is PM10 however the respirable fraction nay be of
importance for the assessment of health effects due to aerosol
exposures from a variety of sources. The flow rates have been chosen
so that they conform to currently available personal sampling pump
capabilities (4 LPH) and to a prototype pump being developed by EPA
capable of drawing 10 LPM. Figure 2 shows a prototype personal
sampling pump disassembled. The components in the foreground are from
left to right, the flow controller, pump and motor, the electronics
circuit board; and battery pack. The background components are the
shell for the flow controller, pump, and motor. This package has been
developed for the EPA by Environmental Monitoring and Services, Inc.
The sample aerosol is drawn through the inlet by the personal pump
and the particles are collected on a teflon filter to be weighed and
preserved for later analysis. Analysis methodology (e.g. XRF) requires
uniform particle distribution on the filter. The impactors have been
calibrated with particles from a vibrating orifice monodisperse
generator. They were found to have sharp cut off characteristics,
low particle interstage losses, and good uniformity of deposition.
A preliminary study has been conducted by the University of North
Carolina.4 Objectives of this study were (1) to evaluate alternative
means for measuring aerosols in homes, (2) to measure the
concentrations and size distributions of particles, (3) to Bake
preliminary judgments about the sources of particles, (4) to determine
the fraction of aerosol mass in fine (2.5 urn) and coarse (2.5-10 urn)
sizes, (5) to measure changes in particle size distributions with
time, (6) to study particle morphology and elemental composition, (7)
and to test new personal sampling inlets. A variety of continuous
sampling instruments were used to discern the size spectra of the
aerosol particles. Microscopic examination of filter samples was mad*
for morphological analysis and elemental analysis. The effective siz*
spectrum of particles from 0.01 to 100 urn aerodynamic particle
diameter was studied. The information collected is being used to help
determine the relationship between indoor, outdoor, and personal
aerosol exposures and to evaluate the sampling equipment for inclusion
in further PTEAM work.
Survey questionnaires are being developed to screen the population
for study selection and to relate activities and sources of aerosol
emissions to increased personal exposures. Information areas to be
surveyed include: demographic information (roster of participant
household, participant occupation, age, smoking status, sex, hobbies,
socioeconomic status, housing type), sources of exposures to aerosols
and chemical species of interest, activities correlating with
exposures, limited health effect or wellness information, ventilation
(air exchange rates,heating and air conditioning sources), residence
descriptives (e.g., multi-unit, attached), transportation (commuting
time, type of vehicle), occupation, and workplace descriptives.
The information will be obtained through the use of several
different questionnaires and forms administered to the participant
86
-------
JU K°r conPleted *>y technical field personnel. The first questionnaire
to t>e used are the household screening questionnaire, which is used to
provide the basic demographic information and any other information
necessary for stratification of the population being sampled. Next
cones the participant questionnaire which is used to obtain household
characteristics, personal characteristics, and workplace
c"a"cterlstics from the subject. A 24 hour activity log is
•aainistered to the participant during the actual day of sampling.
IM activity log will be composed of two parts, (1) a chronological
and (2) • supplemental close-ended questionnaire. Other forms
Telated items to be included in the study are participant
iefing questionnaires to be used in a limited field test (the 9
person pre-pilot study) , a letter of introduction for the interviewer,
* 8*Pl»ining the study to the participant, news releases to
n local interest, a participant results report form, and a home
questionnaire. The home survey questionnaire will be used to
follow"uP information in cases where further investigation is
methods discussed above, a field study is being planned to
pg the aerosol exposure distribution of a metropolitan
inclusions
of J^rticle-TEAM study objectives are four fold. First, an estimate
j>M "e frequency distribution of human exposure to particles in the
PM2.5 »ize ranges will be obtained for a metropolitan
n" Second, the differences among the concentration of
n&tter measured by personal exposure monitoring, outdoor
w r B&npling, and fixed site or microenvironmental monitoring
diseemed. Third, identification of the major sources of
£e and tne de?ree °* exposure of an urban population will be
app ~£ted and used to provide exposure assessment and source
«•».. lonment. Fourth, models for personal exposure and source
s*«»»«nt will be developed.
Vaun A» "The Rol« of Total Exposure Measurement in Risk
Keynote Address" Proceedings of the 1987 EPA/APCA
aurement of Toxic and Related Air Pollutants, pp. 1-4.
*9ane\JV W*' Wallace, L., et al. (1986) The Environmental Protection
12{475-* Researcn Program on Total Human Exposure. Environ. Intern.
3.
)le, v., Liu, B., Behm, S., Olson, B., and Wiener, R.W. (1988)
personal Impactor Sampler Inlet." Presented to the American
Hygiene Conference, San Francisco, CA, May 15-20, 1988.
4
" rSS*1*?' R'' Wiener, R., Lee, C., Leith, D. (1988) "The
^acterization of Aerosols in Residential Environments." Proceedings
of TOXJ
May 2-4, 1988, Raleigh, NC.
87
-------
Figure 1. Photograph of Personal Exposure Monitor (PEM) Inlets, Two
examples using multiple orifices (left) and single orifice (right)
designs. Inlets are disassembled. Each inlet consists of an inpactor
classifier, to remove particles larger than the predetermined cut
size, and -a filter to collect the remaining particles.
Figure 2. Photograph of a disassembled prototype personal sampling
pump. The components in the foreground are from left to right, the
flow controller, pump and motor; the electronics circuit board; and
battery pack, the background components are the shell for the flow
controller, pump, and motor.
88
-------
A PRELIMINARY STUDY TO CHARACTERIZE
PARTICLES IN THREE HOMES
ft3 ajnens, Chung-te Lee,
Russell Wiener*
ener*, and David Leith
Apartment of Environmental Sciences and Engineering
Diversity of North Carolina, Chapel Hill, NC 27599 USA
v£°nmental Monitoring Systems Laboratory
EPA, Research Triangle Park, NC, USA
Thei
°se of this study was to begin the characterization of indoor aerosols. Three rniddle-
omes in Chapel Hill, North Carolina were selected for the study. No occupants of
* were smokers. A single central sampling location between the kitchen and dining
and tw r'Was use(* *n e?c^ of the nomes- ^ & ^ mm prototype personal sampling inlets
conce ° • ** **^10 ambient samplers were used to determine the average paniculate mass
um orV^66 homes we observed that 37% of the particle mass was collected in a fine (2.5
"N fraction, 26% between 2.5 and 10 um, and 37% was in a fraction equal to or
0 um. The particle concentrations obtained with prototype personal samplers
« reasonably well to those obtained with 10 um ambient air dichotomous samplers.
[gnif ~S1ZC *nformation obtained from automated aerosol instruments suggests that the most
acuu1Cant SIn£^ P^ck generating event in all of the households was cooking. Household
• r ^ sweeping was the most significant large particle generating event. Electron photo
.JJsraphs indicated that particles below 1 um dominate the particle size-number
v ution. These observations are supported by samples of the same air taken with
^ated instruments.
89
-------
INTRODUCTION
The US EPA is embarking on a major study to estimate human exposure to particle!
and sources of aerosol matter. The program is entitled "The Particle Total Exposw
Assessment Methodology (P TEAM) Study"(1). Much of the human exposure will tab
place indoors. Small, light-weight, personal samplers have been proposed to estimate fine
particle exposure in the sub 2.5 um and sub 10 um range.
Indoor particles arise from a variety of indoor and outdoor sources. These include
road and soil dust, automobile and other outdoor and indoor combustion soots, pollen, fungal
spores, molds, bacteria, insect feces, dried insect fragments, animal dander, cooking aerosols,
etc.(2) Particles can be directly added to the indoor environment by vacuum sweeping,
dusting, cooking, cigarette smoking, woodstove leaks, ventilation system aerosol
generation.etc. The literature provides almost no information on the size distribution of
indoor aerosols and to what extent different sources contribute to the overall character of
indoor aerosols.
The purpose of this study was to begin the characterization of indoor aerosols. The
main objective was to collect information that would permit generation of particle size
distributions in indoor, residential, non-smoking environments. This included a preliminary
effort to 1) determine the fraction of aerosol mass that typically appears in fine (sub 2.5 um)
and coarse (2.5-10 um) particle fractions, 2) follow changes in particle size distributions with
time as a function of different indoor particle-generation events, 3) collect information on
particle morphology and elemental composition, and 4) compare newly developed personal
sampling inlets with an EPA approved fixed ambient sampler used to collect PMio and
PM2.5 particles. Four personal sampling inlets were made available to UNC for this study by
the Environmental Monitoring Systems Laboratory, U.S.EPA, RTP, NC. These an
described in another paper by Russell Weiner presented at this symposium/1)
APPROACH
Three middle-income, homes were selected for the study, located in Chapel Hill,
North Carolina. No occupants of the homes are smokers. All homes are on wooded lots that
range from 0.4 to 1 acre, are in residential sections of town, and each is adjacent to a quiel
street or lane. Two of the homes are located within a quarter to a half mile of two
thoroughfares that receive heavy traffic in the morning and evening. The third is
approximately one mile from a fairly busy highway. A single central sampling location
between the kitchen and dining room areas was used in each of the homes. At least two 8-
hour sampling periods and one 13-hour evening to early morning sampling period was
conducted in each home.
Standard 47 mm open face samplers, specially designed, prototype 37 & 47 mm
personal sampling inlets, (called Marple inlets because they were designed by V. Marple)
and two fixed PMio ambient samplers were used to determine the average paniculate mass
concentration in each of the homes. The concentrations derived from the PMio and PM2.J
ambient samplers (Sierra Anderson dichotomous samplers with a 10 um inlet) were used as a
comparison base for the personal samplers. Three particle sizing instruments which
collectively spanned the range of 0.01 to 20 um tracked changes in particle sizes during each
sampling period.
RESULTS AND DISCUSSION
Sampling in the homes took place during November and December 1987. The
average temperature in the homes was 21QC and the average relative humidity was 38%,
Outdoor temperatures during the day were 8-10°C lower than indoor temperatures. Air
90
-------
infiltration rates ranged from 0.16 to 0.72 exchanges per hour as determined from the slope
semi-log plots of SFg concentration against time.
. Filter Weighing. 37 mm and 47 Teflon filters (2 um pore size with a polycarbonate
™>g) and 47 mm Teflon impregnated glass fibers (Pallflex #T60A20) were tared on a 0.000
ean-r?torius micro balance (model 4503) in the U.S. EPA balance room. Filters were
^quuibrated for at least 24 hours in the temperature and humidity controlled balance room
20opt(/ we*8hing. The average temperature and the relative humidity in the room were
ft]te '"2 ant* 39%+/-4 respectively. The error associated with the particle mass on 37 mm
confrt w^ich incudes initial and final weighings, was estimated to be +/-9.2 ug (95%
cnn nce level) and the average relative error associated with resulting particle
^ncentrations was +/-17%.
fracf ^oniPa"son °f Sierra and Marple Inlets. The filter mass obtained from the fine
*ith k°m tne two simultaneousty operated Sierra 10 um inlets (see Table 1) agreed to
2Q2\n *r7 the 10 L/min, 10 um personal sampling inlet agreed within 18.5% of one another
samn]SUm °^ ^ Sierra fine and coarse filter masses (relative s.d. ~ 22.5%). Six of the nine
*ere • ^ere within 10%. With the 4 L/min personal sampling inlet, 6 of the nine samples
°btain^i* ^'^ °^tne Sierra coarse and fine filter masses. Similar results (Table 1) were
fract ^e ^ um Personal sampling inlet upon comparison with the Sierra fine
consfa uNineteen percent of all of the Marple filters (both 2.5 and 10 um) showed
fairtiJS1 e disagreement (>35%) with the Sierra inlets. This was attributed to our lack of
T. use ° l^e new personal sampling inlets and in some cases the relatively
Particle masses collected.
fodoo **ar*'c'e Mass Distribution in the Different Size Fractions. The Marple and Sierra
face J. P^icle concentrations were compared to concentrations derived from 47 mm open
SaitU>leH- samP*es- Jf we assume that the 47 mm open face filters collected almost all of the
Pfctjcl Jndoor particle mass, then particle concentration data could be grouped into three
P0r rt. Slze mass fractions: a <2.5 urn fraction, a 2.5-10 um fraction, and a >10 um fraction.
in fmT/~Verage of the three homes we observed that 37% of the particle mass was collected
equal f um or below) fraction, 26% between 2.5 and 10 um, and 37% was in a fraction
t « to or greater than 10 um.
daytim **enerat'°n of Large and Small Particles. The bottom plot in Figure 1 displays the
(house C4inrC)So1 data in *e 2 um to 20 um ran^e on November 20, 1987 from the first home
^schiH samP^- During this day a hot lunch was prepared on the gas kitchen stove,
househ °}j ^ets were vigorously exercised, toast was made and slightly burned, and a total
filter m vacuuming was undertaken. This day was unusual in that 52% of the total particle
fine nSiaPPear?^ in ^e P^le fraction that was greater than 10 um; only 18% was in the
ran8es ° ction. As previously mentioned, the average for all homes in the >10 um
sho\vc th'as ~^^0 an^ m ^e ^-S um» 38%. Inspection of the bottom graph in Figure 2
than iff at rao.st of ^e particle mass in the 2-20 range resided in particles which are greater
u " Um in diameter. The peak in particle mass appears to coincide with the
""ftg/sweeping of the carpets and floors which began downstairs at 1400 hr.
On November 20, 1987, the most significant small particle generating event was the
Ven°n of a lunch for the five research scientists who were in the house at the time. Stir-
^ egetables were cooked on the household gas stove. The EAA number-size particle
t 2\S *** mo.st °f these P^fcs ranged between 0.01 and 0.04 um (bottom graph of
ok ^n,other significant generator of small particles was making slightly burned toast.
C°n^bo ervat*on that cooking food is an important generator of small aerosols was
rated when we sampled in the other houses.
91
-------
Day vs. Night Aerosols. During the daytime, aerosol concentrations appear to be
related to given household events. The top graph in Figure 1 shows the particle volume
(mass) size distributions over time in the 2 to 20 um range for home #1 during the evening
and early morning of November 23-24, 1987. At 2000 hr as a result of cleaning ashes outo!
a wood stove and then starting a fire, the concentration of aerosol mass increased
dramatically. Note that by 2400 hr everyone in the household had gone to sleep and aerosol
concentrations declined. This was most probably due to a lack of household activity and
settling of the large particles. At approximately 700 hr, when members of the household got
up, showered and had breakfast, a rise in aerosol concentrations in all size ranges was
observed. This evening aerosol trend is also apparent in the other households.
Panicles between 1 and 3 um (top graph, Figure 2) for the early morning hours of
November 23, 1987 in house #1 exhibited a pattern of loss which was similar to the larger
aerosols for this evening. The evening behavior for particles less than 1 um on this day did
not follow the same trend as the larger particles (see Figure 2, top graph). After the peak
concentration was reached these small particle levels appeared to remain high and stable
throughout the night. In addition, in all three homes, most of the particles in the 1-3 um
range were below 1 um. Very few particles between 1 and 3 um contributed to the mass
distribution. This strongly suggests that most of the fine panicles which were collected with
either the Sierra-Andersen impactors or the Marple inlets ranged between 0.05 and 1 um.
Particle Morphology and Elemental Analysis. Over 100 photographs of
scanning electron micrographs were taken from 18 Nuclepore filter (0.1 um pore size)
samples. Analysis was performed on a Cambridge model S-200 scanning electron
microscope (SEM). Associated with most of the electron micrographs is an x-ray
fluorescence (XRF) analysis of a particle in the photograph. Based on the elemental analysis
and comparisons with representative analysis and shapes in Volume VI of the McCrone
Particle Atlas,(4) "guided speculations" about the nature of the particles were made. An
example of house dust and road dust from the Particle Atlas indicated that these particles
have a high content of silicon (Si), aluminum (Al), potassium (K), (Ca), etc., content. North
Carolina clay, quartz, feldspar, etc. also have these elements. Particles which exhibited the
elemental patterns shown in Figure 3 were tentatively described as clays, etc. Dander from
human beings and other animals tend to exhibit a "hump or mound" XRF response and do
not show a substantial elemental trace.
The micrographs clearly show that particles below 1 um dominate the number
distribution. Most of these particles are below 0.5 um. In all three homes there are large
numbers of 0.2 to 0.4 um spherical particles. XRF analysis of these particles consistently
shows the lack of typical elements such as Si, Al, K, Ca and Fe, etc. The response is often
mound shape which indicates the presence of carbon. Finally, there were a large number of
agglomerates, which range in size from 0.3 to 1 um and are composed of aggregates of 0.05
to 0.1 um, spheres. These look like photographs we have taken of wood and diesel soot
particles so that we have generically referred to these aerosols as soot particles.
Dusting in house #1 generated large aerosols which ranged from 13 to 100 um
Some of these particles were stove ash, insect parts, and hair. A sample collected in the
vicinity of the toaster while toast was being made had a large number of particles that
resembled soot. Soot particles and salt panicles were collected near an electric stove where
grilled cheese sandwiches were being made. Particles collected at the exhaust of vacuum
cleaners operated in houses #1 and #3 seemed to be mostly mineral in content, but once
again the sample size may not be large enough to distinguish these particles from the general
particle mix.
Of the 95 particles that were individually analyzed, more than 30% were thought to
be of possible biological origin. These included danders, insect parts and irregular shaped
particles with a mound shape XRF analysis. These particles ranged in size from 2 to 70 um,
All three homes had a large number of mineral or clay particles. These included clay, quartz,
magnetite, salt, chalk,, etc. Mineral particles accounted for approximately 30% of the
particles for which we obtained specific XRF information, and spanned a size range similar
92
-------
to the biological particles. Soot particles accounted for 15% of the individual particles
analyzed by SEM and, as mentioned before, these plus the spherical 0.2 to 0.4 um particles
ccounted for most of the particles in the entire size distribution. In general all three homes
Dart? i tO kave tne same kinds °f particles. House #1 may have had a higher percentage of
P«acles assigned to the biological category. House #3 had a higher percentage of analyzed
Jjuneral particles. This seems reasonable since the owners of house #3 do not often allow
eir dog in the house and have a limited amount of carpeting on the floors. In closing,
*ever» it is important to emphasize that 95 electron micrographs and XRF analysis do not
^essaniy represent a good statistical sample of the particles in this three-home study so that
e data must be viewed as being very qualitative.
DEFERENCES
EPvWAiKener' "Measurement and Evaluation of Personal Exposure to Aerosols" 1988
'VAPCA Symposium on Toxic and Related Ah* Pollutants, Raleigh, NC.
Poll • A todoor Air Quality Implementation Plan: Appendix A: Preliminary Indoor Air
"uuaon Information Assessment; (June 1987) U.S. EPA, Office of Research and
vel°Pment, OHEA, ECAO, Research Triangle Park, NC, 277 11.
PmnM*-Kamens' C- L66' D L61*' 0988)i " A Preliminary Study tp Characterize Indoor
#79(33 xln Three Residential Homes" Final Report to PEI, Inc., Cincinnati, Ohio (Contract
"*/•
McCrone, J.A. Brown, I.M. Stewart; (1980), The Particle Atlas. Edition Two;Ann
ss, Ann Arbor.
ACKNOWLEDGEMENTS:
from pS?8 Work was suPPorted by a contract to the University of North Carolina (# 790-87)
Perform 5 Ass°ciates in Cincinnati, Ohio. Scanning electron microscopy work was
auftoR ln Dr' Rptert Bagnell's laboratory in the Department of Pathology at UNC. The
his staff ex^)ress t*le*r app^c^tion for the training and insights provided by Dr. Bagnell and
93
-------
Table 1
Indoor Participate Concentrations Derived from Dlchotomous Samplers,
Marpie Personal Sampling Inlets, and Open Face Samplers
Date*
11-20-870
11-23-87N
11-24-870
11-23-870
12-7-870
12-7-87N
12-08-870
12-10-870
12-10-87N
12-11-870
House
id ft
1
1
1
1
2
2
2
3
3
3
Sierra
1622
2.5 UH
(ug/m3)
11.4
14.9
13.8
9.0
6.9
9.8
9.0
19.1
7.2
6.5
Sierra
«579
2.5 UM
(ug/»3)
4.0
*«
12.6
9.7
6.7
8.9
8.2
18.3
7.8
7.2
S1err
9662
1Qum
(ug/m
to
6
7
7
12
4
9
12
4
6
a
3)
.6
.8
.3
.0
.2
.1
.8
.7
.8
.9
Sierra
9579
lOurn
(ug/ai3)
14.0
**
9.2
4.0
10.5
4.3
7.2
11.8
3.6
7.1
Marple
Mtl
10 un
10l/m1n
(ug/m3)
**
22.6
20.6
11.0
20.4
11.9
8.5
34.7
10.4
18.0
Marple
MB2
2.5 um
10l/m1n
(ug/«3)
18.1
15.2
12.7
2.5
18.3
9.4
9.6
17.8
6.3
7.4
Marple
MI3
10 um
4l/m1n
(ug/m3)
**
22.5
18.7
8.3
20.5
11.9
30.2
36.7
7.9
11.9
Marple
MI7
2.5 UR
4l/m1n
(ug/*3)
**
**
**
**
**
**
**
25.8
5.9
12.9
Total
mass
(ug/m3)
42.0
33.2
32.8
21.4
34.5
20.2
29.2
47.7
14.4
21.9
*N & 0 denote night or day
** Data not available
House II 1s a 1640 ft2 modern, open space, two story frame construction home,
It uses a wood stove as Its primary source of heat, and has a gas furnace and cook stove.
Two adults, one teenager, and two small dogs and a cat live in this house
The address 1s 13 Frances Street, Chapel Hill, NC
House 82 is a 2800 ft2, two story, contemporary 10 year old, frame construction home,
It has a fire place, an electric cook stove and heat pump for heating.
Two adults, a teenager,a fourth grader, and a medium size dog live 1n this house.
The address Is 116 Porter Place, Chapel H111, N.C.
House 83 1s an 1600 ft2 14 year old ranch style house. It has an electric furnace
and electric stove. The heating/air conditioner duct work 1s fitted with a humidifier.
Two adults live in this house. A large dog spends most of Ms time in the back yard.
The address 1s 251 Indian Trail, Chapel Hill, N.C.
94
-------
.
•G;
Figure 2 . Behavior of fine indoor particles. Top graph is observed behavior (PMS data) during
evening-early morning hours; Bottom graph is EAA data taken during the day, large peak at 12:**
due to cooking food.
96
-------
i... E-: :i: r i i HOI icii;
o :i
<> K i;;: v
I
. , i
III i U i
' i ; • ' ^ '
i p H I l ;
' **'
: r. \... i:;:
1 I
x i;;::;;
•<> K I--: <
jf* ^ • Example of scanning electron micrograph and XRF analysis of a clay particle collected
97
-------
ASBESTOS IN RESIDENTIAL ENVIRONMENTS
R. L. Perkins and K. K. Starner
Research Triangle Institute
Research Triange Park, NC
L. E. Sparks
U.S. EPA, Air and Energy Engineering Research Laboratory
Research Triangle Park, NC
There are approximately 102 million housing units in the United State;
with nearly one-half of them being constructed prior to the mid 1970's whe'
the use of asbestos-containing material (ACM) inside buildings decreased
dramatically. The objectives of this study were to determine concentra-
tions of airborne asbestos present in the residential environment and to
compare the Phase Contrast Microscopy (PCM) method of analysis of air sam-
ples with the Transmission Electron Microscopy (TEN) method. Inspection,
bulk sampling, and air monitoring were conducted at six homes with five of
these homes having friable ACM. Bulk and dust (wipe) samples were analyzei
using Polarized Light Microscopy (PLM) and air samples collected side-by-
side were analyzed by PCM and TEM. The highest airborne fiber concentra-
tion as measured by PCM was 0.027 fibers/cc of air, whereas the highest
value of asbestos fiber concentration determined with TEM was 1.253 fibers
cc of air. Analytical results show that TEM is the superior method for
detecting asbestos fibers. Eighty-eight percent of the fibers detected ty
TEM were <5 yirt in length and therefore would not be included in fiber con-
centration determination by PCM.
98
-------
Introduction
The Technical Assistance Program of the Office of Pesticides and Toxic
substances of the U.S. Environmental Protection Agency (EPA) has published
several documents which provide guidance and information on the identifica-
tion of asbestos-containing materials.1-4 These documents adequately
Present background information about the various facets of the asbestos
problem and outline the procedures to be followed in determining if
asDestos-containing material (ACM) is present in buildings, establishing
Pecial operations and maintenance (0 M) programs, and determining the cor-
rect abatement procedures.
Although much of this information is applicable to single family resi-
buiiH** il is directed mainly to schools and other public and commercial
«| 1 dings. Few studies have been directed towards documenting asbestos
Contamination in the home. Burdett and Jaffrey (1986) determined airborne
soestos concentrations in 235 samples obtained in 39 buildings containing
OSDestos materials.5 This study was conducted in the United Kingdom.
There are approximately 102 million housing units in the United
to ia5*6 An estimated 26% were constructed prior to 1940, 40% from 1940
and i ' and 34% Slnce 1970t Tnose nous1n9 units constructed between 1940
mi in would be most I1kely to contain ACM; this means that nearly 40
Dermi nousin9 units could possibly contain asbestos and some 100 million
toe < could be exposed to airborne asbestos fibers. In the EPA's "Asbes-
as 1" Buildings" study (1984) it was estimated that in the United States,
houcl 8> tnere we""6 350,000 residential buildings containing 10 or more
1nnation of airborne asbestos levels, influencing factors such as
condition of ACM, presence of air moving equipment such as fans
in the area containing the ACM, and frequency and type of use
-v.umo *- room> Playroom, storage, etc.) of the area containing ACM were
tors a H to determine cause and effect relationships between these fac-
and airborne concentrations of asbestos.
Selection of Study Residences
reUti«Si(lences cnosen for study were located by communication with friends,
Vfis, and colleagues. The homeowners were assured that all information
99
-------
concerning the locations of homes and names of homeowners would be confide
tial and, although data related to the findings of the study would be pub-
lished, they would not be traceable to a particular residence.
Inspection, bulk sampling, and air monitoring were conducted at six
homes with five of these homes having ACM. In Home A, air monitoring was
conducted prior to, during, and after abatement to remove the ACM. In
Homes B, C, D, and E the ACM has not been removed to date.
A thorough inspection of each residence was conducted and all materi-
als suspected of containing asbestos were sampled for analysis. Dust sara-
pies were also collected in each home. Air monitoring was conducted in the
basement (all visible ACM was located in the basements of all homes in the
study), the living area, and outside of each home site.
Description of Study Residences
In all study homes, the basement served only as a work and/or storage
area. Facilities such as laundries or workshops were common. No home had1
living area (such as playroom) in the basement. Four of the five homes ci»
taming ACM have passive heating systems (radiator) with the remaining hone
(Home A) having a forced air system.
Chrysotile is the only asbestos type noted in the ACM. The largest
portion of the ACM is in the form of asbestos paper with chrysotile content
rS«9ln£ m 35% to 95%* Altnou9h asbestos paper is the most common type t
ACM, the material which displays the highest degree of deterioration is as-
bestos plaster. In all homes this material was very friable.
Methods
Analysis of ACM
Samples of ACM and dust samples were analyzed by polarized light mi-
croscopy (PLM) with dispersion staining using the protocol recommended by
the EPA.o
Air Sampling Strategy
Samples designated for PCM and TEM analysis were collected by the
static (non-aggressive) method. As all areas sampled were in occupied
residences, aggressive sampling was not considered. Sampling sites were
located both in the basement areas and the upper levels of the residences.
These sites were chosen in order to compare airborne asbestos levels of the
area containing the asbestos materials (basement) to those of the upper
living areas of the home. Outdoor air samples were also collected at each
ori^^ TcJ6™111? the leveln?f airborne asbestos in the ambient air. The
PCM and TEM samples were collected side-by-side at each sampling location.
Analysis By Phase Contrast Microscopy
All PCM samples were analyzed in accordance with NIOSH Method No.
7400.9 OSHA uses this technique to measure total airborne fibers in oc-
cupational environments. The EPA lists PCM analysis as one method for
100
-------
airborne fibers as related to satisfactory completion of asbes-
abatement projects.
Analysis By Transmission Electron Microscopy
co A ^e Nuclepore polycarbonate filters were prepared for analysis in ac-
coat HC6 with EPA recommended protocol. The filters were first carbon
tr .and ^en placed in a condensation washer for transfer to the Elec-
tinn ™croscopy grids. It was determined by the analyst that the pondensa-
°n washer is superior to the Jaffe Washer for transfer of filter medium.
tha TEM ana1ysis involved examination of the particulates deposited on
wer Srimpl.e filter Us1n9 a pn111Ps EM40° 120lfv TEM. Asbestos structures
fied s^9nated as fibers, bundles, or matrices and were sized and identi-
qv ^cording to asbestos type. Asbestos type was determined by morpholo-
y and by observing the selected area electron diffraction (SAED) patterns.
Alr Monitoring Results
Overview
OD| concentrations determined by TEM analysis ranged from no detect-
detLTlbers to 1.253 fibers/cc of air. The highest fiber concentration
abat by PCM was 0<027 fibers/cc air- with the exception of the post-
ernent air samples from Home A, fiber concentrations measured by PCM
ni9her in the living area than those concentrations measured in the
lent, although all known ACM was located in the basement area of all
study homes. The living area/basement PCM fiber concentration ratios
jomes B through E were 2.0, 13.5, 3.0, and 5.0, respectively. Higher
Tiber concentrations (nonasbestos) would be expected in the living
ea h
dlso ^ecause of the presence of carpet, furniture, draperies, etc., and
Pared *Cause °^ tne greater occupant activity in the living area as com-
ea to the basement.
dbate borne fiber concentrations as measured by TEM for Homes A (pre-
detenn? ' D and E were n'"gner in the basement than those concentrations
tratio d for tne Tiving area- ^e basement/living area TEM fiber concen-
reUtin ra^os for Homes A» D» and E were 4> 24» and 8> respectively. This
dct1vit Sh"ip did not exist for Homes B and C because of very low occupant
Plinq J£ jn the basement area and unusual conditions present during sam-
n°ne H I "orae C- Asbestos fiber concentrations in ambient air ranged from
vaiue cted to 0.031 fibers/cc of air as determined by TEM. The average
ue was 0.010 fibers/cc of air.
ne following conclusions resulted from this study:
*• Building survey data indicate that a substantial portion of the
housing units in the United States contain ACM, with the most
common type material being pipe, duct, and boiler wrap. This ACM
is commonly friable due to its age (10-50 years), deterioration
from water, and/or physical damage. Thus, the potential exists
for millions of Americans to be exposed to detectable levels of
airborne asbestos fibers.
101
-------
2. Concentrations of airborne asbestos in homes containing ACM may
exceed the OSHA recommended occupational limit of 0.1 fibers/ca
air. A value of 1.253 fibers/cc was determined at study home A;
this is greater than the OSHA limit by a factor of 12.5. It
should be noted that the value of 1.253 fibers/cc included all
asbestos structures observed in the TEN analysis, whereas the OSH
limit considers only those fibers longer than 5 v-m. It should be
emphasized that an air sample represents only a "snapshot" in tii
and the analysis result should not be considered to represent a
constant state condition. Conversely, the study results do
indicate that under certain conditions, airborne asbestos levels
may be elevated above present recommended limits. Burdett and
Jaffrey found in their survey of 39 buildings, that only 20% of
the 235 samples had asbestos fiber concentrations above the limit
of detection of the TEM and the TEM derived concentration of
fibers greater than 5 ym in length exceeded 0.001 fibers/cc at
only one sampling site.5 All preabatement air samples from Home
A exceeded the value of 0.001 fibers/cc when only fibers greater
than 5 \m were considered.
3. The concentration of airborne asbestos fibers in homes having At!
may be directly affected by the following factors:
0 amount of exposed ACM;
0 condition of ACM (friability);
0 type of occupancy use of the area containing ACM;
0 type of heating/cooling system (forced air vs passive); and
0 short term activities such as dusting, and vacuuming may
increase the concentration of airborne asbestos.
4. Fibers <5 urn in length accounted for the greatest number of fit*
determined by TEM analysis (88% of total identified by TEM).
References
1. "Asbestos-Containing Materials in School Buildings: A Guidance Docu-
ment, Part 1," Office of Toxic Substances, U.S. Environmental Protec-
tion Agency, Washington, D.C., 1979.
2. R. N. Sawyer and D. M. Spooner, "Asbestos-Containing Materials in
School Buildings: A Guidance Document, Part 2," Office of Toxic Sub-
stances, U.S. Environmental Protection Agency, Washington, D.C., 191!
3. "Guidance for Controlling Friable Asbestos-Containing Materials in
Buildings," EPA-560/5-83-002, Office of Toxic Substances, U.S.
Environmental Protection Agency, Washington, D.C., 1983.
4. "Guidance for Controlling Asbestos-Containing Materials in Buildings,
Office of Toxic Substances, U.S. Environmental Protection Agency,
Washington, D.C., 1985.
5. G. J. Burdett and S. A. M. T. Jaffrey, "Airborne Asbestos Concentra-
tions in Buildings," Ann. Occup. Hyg., 30:185 (1986).
6. "Annual Housing Survey," Department of Housing and Urban Development,
U.S. Department of Commerce, Washington, D.C., 1983.
102
-------
• Asbestos in Buildings: A National Survey of Asbestos-Containing
Friable Materials," Office of Toxic Substances, l.S. Environmental
Protection Agency, Washington, D.C., 1984.
• "Interim Method for the Determination of Asbestos in Bulk Insulation
Samples," EPA-600/M4-82-020, U.S. Environmental Protection Agency,
Research Triangle Park, N.C., 1982.
1 "Method for Asbestos Fibers. NIOSH Method No. 7400," National
Institute for Occupational Safety and Health, U.S. Department of
"ealth, Education and Welfare, Cincinnati, Ohio, 1984.
103
-------
NITRIC AND NITROUS ACIDS IN
ENVIRONMENTAL TOBACCO SMOKE
Lamb, J. Crawford,
D.J. Eatough, L, Lewis, J.D.
E.A. Lewis and L.D. Hansen
Chemistry Department
Brigham Young University
Provo, Utah 84602 U.S.A.
and
N.L. Eatough
Department of Chemistry
California Polytechnic State University
San Luis Obispo, California 93407 U.S.A.
Nitric and nitrous acids in environmental tobacco smoke have b
determined by sampling with annular diffusion denuder sampling systems
experiments conducted in a 30 m Teflon chamber and in indoor environmen
The effects of residence time and photochemistry on the chemical composit
of both the gas and particulate phase nitrogen oxides in environmen
tobacco smoke have been studied in the chamber. Nitric acid and N02(g) «
minor constituents of environmental tobacco smoke in the chao
environment. The mole ratio of nitrous acid to NO in the chai
experiments was 0.13. In the absence of UV light, this ratio increi
slowly with time. The mole ratio of HN02(g) to N0x(g) in iru
environments was generally comparable to that seen in the chamber stud:
Concentrations of HN03(g) and particulate nitrate in the indoor environs
were always higher than expected from the relative concentrations of NOX
HNOTCg) and particulate nitrate seen in the chamber experiments. Compar
of the relative concentrations of N0x(g), HN02(g), HN03
-------
Introduction
The chemical characterization of nitrogen oxides present in
snvironmental tobacco smoke has mainly been limited to the determination of
^(g)1'^. Few data are available on the concentrations of N02(g) in
invironmental tobacco smoke, and no previous studies on HN02(g) or HN03(g)
In environmental tobacco smoke have been reported. If present, these
compounds would be expected to interact with organic compounds in
Bnvironmental tobacco smoke. The resulting chemistry may be significant
»ith respect to understanding the toxicology of environmental tobacco smoke.
For example, Kamens et al.3-^ have shown that nitrated, mutagenic compounds
are rapidly formed in wood smoke in the presence of 03(g) and N02(g) .
Recent studies have shown more than a ten-fold increase in gas phase
BUtagenicity as a result of this wood smoke chemistry5. The compounds
responsible for this increase in mutagenicity are very labile and have not
yet been identified. The gas phase mutagens appear to result from the
reaction of nitrogen oxides with organic compounds. If environmental
tobacco smoke contains significant quantities of N02(g), HN02(g) and/or
HN03(g), similar chemistry could also result in the formation of nitrated
organic compounds. The only major class of potentially mutagenic oxy-
nitrogen compounds in environmental tobacco smoke for which data exist are
the N-nitrosamines6'9. These compounds have been shown to be present in
increased concentrations in environmental tobacco smoke as compared to
tobacco smoke condensate, but it is not known if the N-nitrosamines are
formed only during the combustion of the cigarette or are also formed as
secondary products in the atmosphere.
The interaction of organic compounds in environmental tobacco smoke with
nitrogen oxides and oxy acids to form new compounds will occur only if
compounds other than NO are present at significant concentrations in indoor
atmospheres. Results from studies to determine the concentrations of
U02(g), HN02(g) and HN03(g) in environmental tobacco smoke are reported
here.
Experimental Methods
Environmental Chamber Studies. The 30 m3 Teflon chamber, associated
equipment, and the diffusion denuders used for the selective collection of
gases in the presence of particles have been described1"-•"">. Details of the
:experimental techniques used for the determination of the organic components
of environmental tobacco smoke in the chamber and indoor studies are given
elsewhere12'14.
Sampling. The acidic gases, HN03(g), HN02(g) and S02(g), were
collected with NaHC03 coated annular diffusion denuders11. The denuder
surface was prepared using a 1 wt% NaHC03/glycerine aqueous solution by
twice wetting the denuder surface with about 5 mL of the solution, draining
the excess solution and drying the coating with an N2 stream. The diffusion
denuder sampling device consisted of a Teflon cyclone to remove particles
larger than 2 fim, two coated annular denuder sections, a 47 mm Teflon
nembrane filter and a 47 mm Nylon membrane filter, in the order listed.
Sample flow rate through the system was controlled at about 15 slpm using a
Tylan mass flow controller. The concentrations of C0(g), N0(g), N0x(g) and
SOo(g) were determined in the chamber experiments with real-time monitoring
instruments10'12. In a few experiments in the chamber, the concentrations
of N02 were determined using a Scientrex Luminol N02 analyzer. The
concentrations of N0x(g) and C0(g) in the indoor environments were obtained
using Drager tubes. The flow rates for all sampling systems except the
Drager tubes were controlled with Tylan mass flow controllers. The Tylan
105
-------
units were calibrated against a Kurz mass flow meter and a dry gas mete:
The sample flow rates for the Drager tubes were controlled by a critic
orifice and monitored with rotometers calibrated against a bubble fl:
meter. The Drager tubes (Dragerwerk A.G., Lubeck) were calibrated agair:
the real-time C0(g), N0(g) and N0x(g) instruments during sampling fromd
environmental chamber.
Analytical Techniques. Water soluble anions from gases and particle
collected by the annular diffusion denuder were determined by it
chromatography. The material collected in the denuder sections and on ft
Teflon filters was extracted with water. The Nylon filters were extract!;
with the Na2C03/NaHCC>3 ion chromatographic eluent. The identification:
nitrite in the extract solutions was confirmed by the use of bot
Na2CC>3/NaHC03 and NaHC03 eluents and by spiking the samples with nitrit
standards. All analytical results were corrected for blank concentration;
Blank samples were handled and analyzed the same way as the environments
samples except that no air was passed through the blanks. The data:
nitrate and nitrite in the first annular denuder section were corrected f:
the partial absorption of gases such as N(>2(g) using the nitrate and nitrit;
concentrations found in the second denuder section^- .
Results
Chamber Experiments. The gas and particulate phase concentrations t
nicotine, 3-ethenylpyridine, and inorganic nitrogen and sulfur oxidt
relative to C0(g) in environmental tobacco smoke in the environment!
chamber are given in Table I. The chemical composition of ETS in t!
chamber was determined for samples collected immediately after combustion::
the cigarettes and after two to four hours of aging. Samples were collect!!
for environmental tobacco smoke produced from the combustion of 1, 2, 3 at)
4 cigarettes. The gas/particle distributions of the compounds listed!:
Table I did not vary substantially with time or number of cigarettes smoked
In addition, the concentrations of nicotine, C0(g) and N0x(g) were constarj
with time in the chamber as indicated by the data in Figure 1. Hi
concentration of HN02(g) and the difference between the N0x(g) and N0([!
concentrations measured with the Monitor Labs Model 8840 Nitrogen Oxidsj
Analyzer increased slowly with time as shown in Figure 1. The different!
between the measured N0x(g) and N0(g) concentrations may be due to respons
of the instrument to HN02(g). Experiments using the Scientrex Luminol tf
instrument indicated that the concentrations of N02(g) are much lower the
inferred with the measurements using the Monitor Labs instrument, Table :
The combined data suggest that the only major inorganic nitrogen oxia
present in freshly produced environmental tobacco smoke other than NO ii
HN02(g).
Significant chemical changes occur when the environmental tobacco smofe
in the chamber is exposed to 385nm UV light as illustrated in Figure 2.
decrease in gas phase hydrocarbons and an increase in particulate mas
results from the exposure to UV light. The decrease in gas phase nicotic
is due to both the formation of particulate phase nicotine and th
conversion of nicotine to other organic compounds^,13 As expected, ft
radiation results in the photoxidation of N0(g) and an increase in tt
concentration of N02(g). The N02(g) reacts to form other compounds, £
indicated both by the Lurainol N02 measurements and by the (N0x(g) - N0(g,'
concentrations obtained with the Monitor Labs instrument. The possibl:
production of HN02(g) or HNC>3(g) was not investigated in these experiments,
Indoor Experiments. A description of the indoor environments studied at
the concentrations of C0(g), N0x(g) and suspended fine particles seen intb
106
-------
ndoor environments are given in Table II. The concentrations of gas and
jrticulate phase nicotine, gas phase 3-ethenylpyridine and the inorganic
Itrogen and sulfur oxides seen in the indoor environments are given in
able III. Nicotine and 3-ethenylpyridine were not seen in the environment
Ith a wood fireplace and no environmental tobacco smoke present. Both
^(g) and HN032(g) and the sum of HN02(g), HNC>3(g) and
articulate nitrate compared to N0x(g) in the indoor environments and
ipected from the chamber studies are given in Figure 3. The fraction of
jtal inorganic nitrogen oxides present as nitrate and nitrite species in
ie indoor environments is comparable to, or slightly higher than that seen
n the chamber experiments .
The concentrations of HNC>2(g) compared to the concentrations of 3-
thenylpyridine, a conservative tracer of environmental tobacco smoke1^'1-1,
i all the indoor environments with ETS except for the disco are given in
Igure 4. The data suggest that the concentrations of HN02(g) in these
idoor environments is equal to, or greater than that expected based on the
ssults of the chamber experiments. The concentrations of HNC>3(g) seen in
ie indoor environments, Table II, were always higher than that expected
rom the chamber experiments, Table I, by at least an order of magnitude.
le mole ratio of the sum of the concentrations of HNC^g), HNC>3(g) and
irticulate nitrate in the disco to particulate nicotine and 3-
thenylpyridine, 14.5 and 2.7, respectively, are comparable to corresponding
itios of 12.3 and 3.5 found for environmental tobacco smoke in the chamber
tudies, Table I. This suggests that HN023(g) in indoor environments.
The increased concentrations of HN023(g), may be formed as the tobacco smoke
ges, as suggested by the data in Figure 1. HNC>3(g) may be formed from the
N02(g) in environmental tobacco smoke. HN02(g) and HNC>3(g) may be present
ron other combustion sources. The contribution of environmental tobacco
noke to the total HNC>2(g) and HN03
-------
References
1. "The Health Consequences of Involuntary Smoking," U.S. Department
Health and Human Services, Washington, DC, 1986.
2. "Environmental Tobacco Smoke. Measuring Exposure and Assessing Heal
Effects," National Academy of Sciences, Washington, DC, 1986.
3. R.M. Kamens, D.A. Bell, A. Dietrich, J.M. Perry, R.G. Goodman, L
Claxton, S. Tejada, "Mutagenic transformations of dilute wood si
systems in the presence of ozone and nitrogen dioxide. Analysis
selected high-pressure liquid chromatography fractions from wood sn
particle extracts," Environ. Sci. Technol. 19: 63-69 (1985).
4. R.M. Kamens, G.D. Rives, J.M. Perry, D.A. Bell, R.F. Paylor Jr., L
Goodman, L.D. Claxton, "Mutagenic changes in dilute wood smoke as
ages and reacts with ozone and nitrogen dioxide: an outdoor chani
study," Environ. Set. Technol. 18: 523-530 (1984).
5. L.T. Cupitt, L.D. Claxton, P.B. Shepson, T.E. Kleindienst, "It
emissions: transformations and fate," Proceedings. EPA/APCA Svmposi
on Measurement of Toxic and Related Air Pollutants. Air Polluti
Control Association, 597-604 (1987).
6. J.D. Adams, K.J. O'Mara-Adams, D. Hoffmann, "Toxic and carcinoget
agents in undiluted mainstream smoke and sidestream smoke of differs
types of cigarettes," Carcinogenesis 8: 729-731 (1987).
7. K.D. Brunnemann, L. Genoble, D. Hoffmann, "Identification and analyi
of a new tobacco-specific N-nitrosamine, 4-(methylnitrosamino)-4-
pyridyl)-l-butanol," Careinogenesis 8: 465-469 (1987).
8. D. Hoffmann, J.D. Adams, K.D. Brunnemann, "A critical look at
nitrosamines in environmental tobacco smoke," Toxicology Letters 35:
8 (1987).
9. G.Stehlik, 0. Richter 0., H. Altmann, "Concentration
dimethylnitrosamine in the air of smoke-filled rooms," Ecotoxlcol
and Environmental Safety 6: 495-500 (1982).
10. D.J. Eatough, C.L. Benner, R.L. Mooney, D, Bartholomew, D.S. Stein
L.D. Hansen, J.D. Lamb, E.A. Lewis, "Gas and particle phase nicotine
environmental tobacco smoke," Proceedings. 79th Annual Meeting of
Air Pollut. Contr. Assoc. Paper 86-68.5, 22-27 June, Minneapolis,
(1986).
11. N.L. Eatough, S. -McGregor, E.A. Lewis, D.J. Eatough, A.A. Huang, E
Ellis, "Comparison of six denuder methods and a filter pack for
collection of ambient HN03(g) and HN02(g) in the 1985 NSMC stuii
Atmos. Environ., in press (1988).
12. D.J. Eatough, C.L. Benner, J.M. Bayona, F.M. Caka, G. Richards, J
Lamb, E.A. Lewis, L.D. Hansen, "The chemical composition
environmental tobacco smoke. I. Gas phase acids and bases," Envii
Sci. Technol.. submitted (1988).
13. C.L. Benner, J.M. Bayona, F.M. Caka, H. Tang, L. Lewis, J. Crawf<
J.D. Lamb, M.L. Lee, E.A. Lewis, L.D. Hansen, D.J. Eatough, '
chemical composition of environmental tobacco smoke. II. Particu]
phase," Environ. Sci. Technol.. submitted (1988).
14. D.J. Eatough, C.L. Benner, H. Tang, V. Landon, G. Richards, F.M. Ci
J. Crawford, E.A. Lewis, L.D. Hansen, N.L. Eatough, "The chem:
composition of environmental tobacco smoke. III. Identification
conservative tracers of environmental tobacco smoke," Environ. Infr
submitted (1988).
15. D.J. Eatough, C.L. Benner, J.M. Bayona, F.M. Caka, H. Tang, L. Le<
J.D. Lamb, M.L. Lee, E.A. Lewis, L.D. Hansen, "Sampling for gas
particle phase nicotine in environmental tobacco smoke with a diffu
denuder and a passive sampler," Proceedings. EPA/APCA Symposiun
Measurement of Toxic and Related Air Pollutants. Air Pollution Con
Association, 132-139 (1987).
108
-------
En i I- Selected Gas and Particulate Phase Compounds Identified in
"wironmental Tobacco Smoke Equilibrated in a 30 m3 Teflon Chamber.a
Chemical
Conroounfl
Bases
In°rganic
Immediately after Combustion After 3 Hours
/*Mol Compound/Hoi CO /iMol Compound/Mol CO
Gas Phase Particles Gas Phase Particles
Nicotine
3-Ethenyl-
pyridine
N0(g)
N02(g)b
N02(g)C
HN03
HN02
S02 + Sulfate
1250013600
1370+460
3670013200
2520+ 430
101 5
961 127
465011300
711 52
3931150 9900
<5 1200
0
0
0
921 55
54+ 40 752011500
71+ 40
<5
30115
4.2210.82g/molCO
from ref. 12, 13, and 14.
j.2 determined from the difference between NOX and NO measured with a
c No tor ^bs Model 8840 Nitrogen Oxides Analyzer.
2 Determined using the Scientrex Luminol instrument.
TT
Concentrations of CO, NOX and Particles (<2/un size) in Indoor
r°nments.
Hom6
Day
Jay 2
Replicate
Site Samples
Living Room 2
Dining Room 1
Living Room 2
Dining Room 1
Break Area 2
Number of ?**£$
Smokers CO. ppm NOX. ppb /ig/m-
0, Wood Burning 0.1910.01 510 151
Fireplace
2, Moderate
Smoking
1, Infrequent
Smoking
2, Moderate
Smoking
1, Infrequent
Smoking
0.84 4 68
0.6710.10 811 35+5
1.90 18 36
0.2710.06 711 2910
Office
Lunchroom
Library
Office
Lunchroom
Library
By Stage
1
1
1
1
1
1
1
1, Heavy Smoking 3.58
in A.M.
2-5, During 2.17
Breaks
0 1.21
1, Heavy Smoking 2.80
in A.M.
2-4, During 2.59
Breaks
0 2.49
20-100 20.512.3
109
42
42
27
51
53
44
131+24
28
94
31
52
76
38
788+19
-------
Table III. Concentrations of Nicotine, 3-Ethenylpyridine and Inorganic Nitrogen and Sulfur Oxides in
Environmental Tobacco Smoke for Samples Collected in Indoor Environments. The locations are the Same
as Those Detailed in Table II.
Chemical Components of Environmental Tobacco Smoke, nmol/m3
Home V13
Home S
Home L
Hair
Salon
Business,
Day 1
Business,
Day 2
Disco
a3-EtyPyr
ULT^.
-------
(0
c
ca
*-
c
Q>
o
c
o
o
0)
• ^^
+••
a
"55
DC
10
9
8
7
6
5
4
3
2
Hydrocarbons* 4, ppm
Nicotine*? 0
HNO2-30
100
140
180
Pi
Time After Combustion, minutes
L Changes in the concentrations of gas phase hydrocarbons, C0(g),
—• (N0x(g)-N0(g))( HN(>2(g), gas phase nicotine and particulate mass
-nS aging of environmental tobacco smoke in a 30 m3 Teflon chamber.
CO
C
o
0)
o
c
o
o
0)
QC
.' Hydrocarbons* 2
-10
30
70
110
150
190
p
Time After Combustion, minutes
Changes in the concentrations of gas phase hydrocarbons, C0(g),
-------
[HNOx T]
[HN02(g)l
"o
c
03
O
X
O
I
450
400
350
300
250
200
150
100
50
0
Predicted from *
Chamber Data /f '.
s''''
• - " X'
* • ,S'
• " * s*'
* X*
•» '* *
* s
4OU
400
350
300
250
200
150
100
50
0
n
E
0
c
0
*w
CM
0
I
1000
2000
3000
[NOx(g)l, nmol/m3
3. Concentrations of N0x(g). HN02(g) and HNOX Total (the sun
HNO,(g) and particulate nitrate) in indoor environments
nvironfal tobacco Loke present. The solid lines are the val
predicted from the results of the chamber experiments.
CO
E
"o
c
[HNO2(g)
250
200
150
i
100
50
0
<
Predicted from
Chamber Data
a NV
a a ..---"'
a ^"'
s--"" ' D
B*-"
0 10 20 30 40 5
[3-Ethenylpyridine], nmol/m3
Figure 4. Concentrations of HN02(g) vs 3-ethenylpyridine in W
environments with environmental tobacco smoke present. The solid line
that predicted from the results of the chamber experiments.
112
-------
r PERSONAL AIR SAMPLING
COMPONENT OF TOTAL HUMAN EXPOSURE
Buckley, J.M. Waldman, and P.J. Lioy
Department °f Environmental & Community Medicine and
6*" Environraental Science of Rutgers University
675 nl ert Wood Johnson Medical School
S Lane
, NJ
A successful approach for sampling personal air as a component of the
total human community exposure is presented. The study
Thig j e1uires 24-h personal sampling at 4 1pm for 14 consecutive days.
sien fthflii«x« sampling procedures and instrumentation
einployed in occupational or environmental sampling. Personal
Cons^
-------
INTRODUCTION
This paper outlines a successful approach for managing high-flow, 24-
hour continuous community personal sampling. The personal sampling
strategies discussed in this paper were developed to meet the requirements
of a study designed to measure total human exposure to benzo[ajpyrene
(BaP). This study, the Total Human Environmental Exposure Study (THEES)^
is a four year study sponsored by the New Jersey Department of
Environmental Protection. As a product of combustion, BaP is ubiquitous i:
the environment and is found adsorbed to particles. As a component of
THEES, the personal sampling discussed here is designed to characterize tin
inhalation exposure. This paper does not present data from THEES, but
rather the successful method by which inhalation exposure samples were
collected via personal samplers.
To measure exposure due to inhalation, fixed station samplers (e.g.,
indoor, outdoor) have traditionally been employed. More recently, personal
air sampling has increasingly become a desirable method of assessing human
inhalation exposure.^ Relative to fixed station samplers, personal
samplers provide a more representative sample of what a person is actually
breathing3''* and provides samples more closely associated with health
outcomes. •*
To conduct the high-flow, 24-hour continuous sampling required in thi.1
study, it was necessary to develop a means of powering the pump for 24-h(K
operation and to devise a way of reducing the pump sound level.
AN APPROACH TO PERSONAL SAMPLING
The personal sampler consisted of a high-flow personal pump (DuPont,
Model P4LA) and a personal impactor (designed by Virgil Marple, University
of Minnesota) loaded with a 25 mm glass fiber filter. The sampling
instrumentation was enhanced for community application with a carrying cast
and an acoustic shell. The personal pump purchased for this study was
selected based on its ability to maintain constant flow rate of 4 1pm at a
pressure drop of >10 inches of water; longevity of operation from a single
battery charge; the ability to operate from both battery and AC current;
and lapse time readout.
The carrying case (Gilian) was selected because of design features
which included both a shoulder and waist strap, providing maximum support,
comfort and versatility. The total weight of the sampling apparatus was
6.0 Ibs, measuring 8" x 4.25" x 6.5".
The study was designed to collect particles with a mass median
aerodynamic diameter <10 urn (PM-10) at a flow rate of 4 1pm. This protocol
provided adequate sample for the analytical detection of BaP.^ Sampling
periods were of 24-hour duration for 14 consecutive days.
Pump Sound Mitigation
At 4 1pm, and a pressure drop of 8 inches of water, the pumps operate:
at approximately 75 dBA (measured at a distance of 12 inches) . This level
of sound exceeds most ambient and indoor background sound levels and would
not be tolerated on a continual basis by the study participants. To
accomplish high-flow personal sampling, a significant reduction in sound
level was imperative.
Investigating and developing a means of sound mitigation was a multi-
faceted process. Initial efforts involved modifying the motor mount,
114
-------
ancement of sound absorbing material around the pump, and muffling the
ound transmitted through the exhaust and inlet ports of the pump. These
rategies were not found to be especially effective, with sound level
Auctions of only 1-3 dBA.
s^ A more successful strategy entailed enclosing the pump in an acoustic
1- To meet the aesthetic and sound mitigation criteria, a 1/4 inch
b_ stical foam barrier composite consisting of a dense loaded vinyl
dec r d lb/sq ft) and a polyester open cell acoustical grade foam
Tee>,U 6r ^ lbs/cu ft) was selected. This material is manufactured by
tjj nicon (Concord, North Carolina) and was applied as a bilayer, covering
all e and outside of an aluminum frame. The shell was constructed to
ei easy access to the pump for daily maintenance and for an external
thaf. Cal connection occasionally used to power the pump. It was observed
eric, fc"ere was no appreciable heat or pressure build-up within the
55 d_ n the average, the acoustic shell resulted in a reduction from 75 to
?" For situations requiring additional sound dampening (e.g. ,
°r readinS> a small duffel bag, lined with a single layer of the
reduc ;°ain composite, was provided resulting in an additional 7 dBA
\f
ng the Power Requirements
P11™?13 were purchased with the manufacturer's 5-cell nickel/cadmium
gr^' This battery is rechargeable and has a power rating of 2.2 amp
8 h0 ' T*le manufacturer's battery design for industrial hygiene allows for
of w *s °^ high- flow operation. At 4 1pm and a pressure drop of 10 inches
opet ®r> *-fc was confirmed that the battery provides at least 8 hours of
houts even under heavy filter loading. To provide the power for 24
batter Continuous operation, a combination of methods were employed:
house ^ cycling (2 batteries for 16 hours) and operating the pump from
oper at?Urrent (8 hours) with a 12 volt charger adaptor. The house current
Were °n of the pump was employed during the sleeping hours. Batteries
^ejtibi?d ^Urin8 tlie course of the day and evening when portability and
lUty Were most important.
TVi
To acc recllar§e cycle for the nickel/cadmium batteries required 14 hours.
Vete Ommodate the pump powering and battery recharging cycle, 3 batteries
er pumP- ^° Prevent the development of "memory," the
neec*e<* to De fully discharged prior to each charging cycle. The
Were fully discharged by placing a 10 Ohm resistor across the
ar Until ttie voltage dropped below an established level. The
rSe cycle required < 2 hours.
Tli
0 P°wering schedule required two maintenance visits per day: one in
rrio n8 before the participant left for work and the second in the late
°n at which time the filters were changed and the flow calibrated.
xi requtred in the field for each of these maintenance visits was
"lately 3 and 10 minutes respectively.
Study
Th
^ Participants were a diverse group with 8 women and 6 men (there
les) , ranging in age from 29 - 79 with a median age of 42. The
. of these people were varied: 3 participants were retired and
ft mo,st all of their time at home; 10 participants were employed
°f the home.
115
-------
Participants were asked to have the sampler near them at all times
to be worn whenever convenient. At night the sampler was placed on the
bedstand near the bed. Participants were advised that the personal s8!$
was not to prevent them from their routine activities. The personal
sampling protocol required only passive involvement of the study
participants, i.e., participants only had to wear the sampler. All
maintenance tasks (battery and filter changes, turning the pump "on" a$
"off") were performed by field technicians. Technicians were on call &•
h/day to respond to any problems with the sampler.
During the 194 person-days of the study, approximately 50 samples
(26%) were interrupted for various reasons. These interruptions were
generally of short duration (2-3 h) and resulted from 3 sources: 1)
insufficient battery charge, 2) pump malfunction, and 3) participant
interference (e.g., obstruction of the inlet or kinking of the tubing)'
Overall, only 1 of 194 (0.5%) person-days was entirely lost.
3 samples had 48 hour run times due to the unavailability of the
participant for the filter change. This low figure is attributable to
high degree of participant commitment and cooperation, 24-hour availabt]
of researchers , and a sampling protocol which was planned and practiced;
VERIFICATION OF THE PERSONAL IMPACTOR PM-10
The personal impactor selected for personal sampling had not been
before. Therefore, a performance test was conducted to demonstrate
variability of PM-10 mass measurements among the Marple personal
prior to the field application, A second intercomparison study was
conducted to verify the PM-10 mass measurements of the Marple personal
impactor against other PM-10 samplers.
The performance test compared the PM-10 data for five side-by-sidfl
Harple personal Impactors. 24 hour samples were collected at 4 Iprn fc*
four days, This test showed a great deal of variability in mass
concentrations between the samplers within a given day with an average
coefficient of variation of 1.38 for the 4 days of samples. It was nOfi
during the sampling that a likely source of this variability was the
adhesion of the glass -fiber filter to the sllicone gasket. Consequent
the silicone gasket was replaced with a Teflon gasket. No adhesion w*1
observed to occur subsequent to this replacement. The results from tW
test conducted using the Teflon gasket are in sharp contrast to the *•*
of the test when the silicone gasket was used (coefficient of variatio13
0.032 vs 1.38).
To verify the particulate mass cut size of the new PM-10 Marple
personal impactor, three types of PM-10 samplers were compared in a f $
test. This test compared the mass collected with the personal impact^
1pm) against the Indoor Air Sampling Impactor (IASI)7 (10 urn cut at 1<>
1pm), and the Sierra dichotomous sampler (10 urn cut at 16.7 1pm) . In*
recent study at our laboratory8, the PM-10 data from the IASI and the
dichotomous sampler were compared and produced excellent results.
The three sampler types (5 personal impactors, 3 lASIs and 1
dichotomous} were compared by simultaneously measuring the PM-10
concentration. The samplers were run in close proximity to one anothe*
space measuring 6 x 2.5 x 2.5 m. Samples were collected on glass flb«*
filters: 25, 37, and 47 mm diameter, respectively. The dichotomous 6
indoor samplers were oriented In an upright position and the personal
samplers were suspended in a fashion similar to being worn. A fan
away from the monitors assured a well -mixed aerosol distribution.
116
-------
mas ?S W6re collected for 5 days- To test the PM-10 cut over a range of
s loadings, the particulate concentration in the sample room was varied,
50 ug/m3.
With slopes and correlation coefficients close to 1, the relationships
;trated in Figure 1 suggest that, within the limits of this experiment,
c Personal sampler provides an accurate estimate of the PM-10
Ca entration. The filter cassettes used with the dichotomous sampler
PM ?n S0me filter shearing and may have added to the variability of its
~i0 determination.
gure 1.
Relationship of the Personal Sampler PM-10 to that of the
IASI (left) and the Dichotomous Sampler (right) .
00-
lo-
- 0.98
- 0,97
w-io
M-IO.
40
On/"1)
a method of measuring human exposure, personal sampling is evolving
-.,« a terms of instrumentation and application. This study demonstrates
hour ^plication and some of the obstacles in implementing high-flow, 24-
Partic?nt*nuous environmental sampling. In this paper, pump sound level,
Pant cooperation, and powering of the sampler are addressed.
"^P sound level was found to be most effectively controlled by
the pump in an acoustic shell. This shell was designed to
the added weight and volume of the sampler. The sampler operated
PattjZr'" •* d^A in the shell. This sound level was acceptable to all
tesluir jSnts although, in some cases, additional sound mitigation was
at
sleeping. In these cases a duffel bag insulated with
ttater*al was employed. The personal sampler was susceptible to
e «ion and» therefore, successful continuous sampling required a high
Petson i Participant cooperation and researcher availability. The
°f bat-*. SamPler was operated 24 hours continuously by using a combination
8Uffj ery cycling and house current. This strategy was designed to assure
n* P°wer and minimize participant inconvenience. Experience with
Personal imp actor showed that it performed most reliably when used
^°n gasket rather than silicone (when used with a glass fiber
At* intercomparison study provided data supporting the PM-10 cut
of the personal impactor.
needs for the development of a personal sampler suitable for
sampling is apparent from the personal sampling conducted in this
ori6 is recommended that such a sampler be designed in three separate
nta: motor and pump; battery; and electronics. Such a design would
117
-------
be advantageous In: isolating the motor and pump for acoustic insulati*
more evenly distributing the weight of the sampler; providing easier *<
to the battery for removal and replacement; and variously designed
components could be interchanged to suit the specific sampling
requirements.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the able assistance of Natilie
Freeman, John Konczyk, Ramana and Rosaline Dhara in the manufacture of
acoustical shell used in this study. Carla Buckley is acknowledged fo*
editing assistance. Funding for THEES is provided by the New Jersey
Department of Environmental Protection, Office of Science and Research-
Mr. Buckley received fellowship support from the Environmental and ;
Occupational Health Science Institute during the year in which this s&
was conducted.
REFERENCES
1. Lioy, P.J., Waldman, J., Greenberg, A., Harkov, R. , and Pietari*1
C., "The total human environmental exposure study (THEES) to
benzo[a]pyrene: Comparison of the inhalation and food pathways)
Arch. Environ. Health, (in press 1988).
2. Wallace, L.A. and Ott, W.R., "Personal monitors: a state-of-th«'
survey," J. Air Pollut. Control Assoc. 32(6): 601 (1982).
3. Cortese, A.D. and Spengler J.D., "Ability of fixed monitoring
stations to represent personal carbon monoxide exposure,"
Pollut. Control Assoc. 26(12): 1144 (1976),
4. Dockery, D.W. and Spengler, J.D., "Personal exposure to respi**B
particulates and sulfates," J. Air Pollut. Control Assoc. 31(2)
(1981).
5. Silverman, F., Corey, P., Mintz, S., Olver, P., and Hosein, R-*
study of effects of ambient urban air pollution using personal
samplers; & preliminary report," Environ. Intern.. 8: 311 (19
-------
Cow AND CHEMICAL CHARACTERIZATION OF WORKPLACE ATMOSPHERES
LAMINATED WITH AIRBORNE DIESEL EXHAUST*
R. A
R, H' ^enKins, W. H. Griest, J. H. Moneyhun, B. A. Tomkins ,
• , C. E. Higgins, and T. M. Gayle
vM T
ru Chemistry Division
Rrjentation and Controls Division
Oak R7?ge Nat*-onal Laboratory
Rldge, Tennessee 37831-6120
^°tenti i^°SUre to aSed diesel exhaust in enclosed workplaces may pose a
is aval] tlealth risk to vehicle maintenance personnel. Only limited data
i * 6 °n t*le influence of dilution and aging on the composition of
larSe dl-esel engines. In order to describe the magnitude and
°f exposure to exhaust- related volatile and less volatile
Such as polynuclear aromatic hydrocarbons (PAH) in the workplace,
*Ve sampling effort was conducted inside selected motor pool
e es at Ft. Carson, Colorado. Samples of fresh exhaust from idling
&CS> tne major contributor to the workplace air contamination, were
ng pp ed* Analyses of vapor and particle phase samples were performed
centr HPLC> and GC/MS. Total suspended particulate (TSP)
to ati°ns ranged from 12 to 300 ug/m3, and were bimodally distributed
size ^mass median diameters of 0.45 and 3.5 microns). The
°mat°g*aphable organic vapor phase compounds were a series of n-
m C5 to C16' Denzene> and alkyl benzenes. PAH's were relatively
E/^"!*1 ^U fc^e Particulate phase. Levels of benzo[a]pyrene ranged up
v°t a c ' The results suggest that fresh exhaust from idling engines is
alth rl°I?p0sitionally accurate surrogate for workplace atmospheres in
sk assessment studies.
"
Sel °nsiderable questions remain regarding the potential toxicity of
exhaust to humans, although the presence in diesel engine
toxic, mutagenic, and carcinogenic compounds is well-
And while questions exist regarding potential occupational
ca by the U' S' Department of Defense and the U. S. Army
No. . ^esearch and Development Laboratory under Interagency Agreement
np 1464"A1 with Martin Marietta Energy Systems, Inc., under
"E-AC05-840R21400 with the U. S. Department of Energy.
119
-------
exposure to airborne diesel exhaust constituents, there exists littlt
direct information concerning the magnitude and composition of sue:
exposures, especially when the likely contributors are large diest
engines. The work here describes an effort to characterize the organi;
chemical composition of the workplace atmosphere in which militar
personnel are exposed to current, petroleum-derived diesel exhaust relate:
contaminants the most frequently and at the highest concentrations, Thi;
objective included a comparison with the composition of diesel engiis
exhaust as the major suspected contributor to the contamination of thi
workplace atmosphere.
EXPERIMENTAL
Sampling was conducted at Ft. Carson, Colorado during two trips madi
approximately two years apart. The primary focus of the first trip wast
obtain samples of exhaust from idling diesel engines of a military nature
Exhaust samples were collected from a wide variety of military vehicles, as
well as diluted dynamometer test stand exhaust from similar engines.
Limited workplace samples also were acquired. The focus of the second tri;
was the collection of samples of workplace atmospheres which had bee:
observed to be contaminated with diesel exhaust. Workplaces included mote:
pool garages and vehicle overhaul areas.
Sampling of the vehicle exhausts was conducted at the individua:
motor pools where the vehicles were located. Sampling of the dynamomete:
exhaust was conducted at the exhaust stacks of the test stands. Flexibls
aluminum tubing was used to channel the diluted exhaust to the sampling
equipment. Typically, the vehicle would be maintained at idling speed for
one hour. Outside air was mixed with the exhaust such that at no time dit
the temperature of the stream actually being collected exceed 52°C.
Workplace air samples were acquired using the sampling equipment
described below at several locations within the confines of a number oi
motor pool areas. Both time resolved (TR) and time averaged (TA) samples
were collected at multiple locations in these facilities to allow at
assessment of the temporal and spatial variability of the workplace
atmospheric composition. Background samples were acquired outside the
buildings where the workplace air samples were taken.
For long term sampling and collection of volatile constituents a
large Tenax trap previously constructed at ORNL for the source sampling of
a coal gasifier1 was employed. The organic compounds collected were
determined by thermal desorption capillary column GC. A portion of the
homogenized Tenax unloaded from a trap was thermally desorbed, and the
compounds liberated were cryogenically focused at the head of the capillary
column before the column oven was temperature programmed. The trapping
system employed for the time resolved collection of organic vapor phase
constituents was a triple sorbent trap developed at ORNL2.
A small low volume cascade impactor (IN-TOX Type 02-100 Mercer
Impactor, IN-TOX Products, Albuquerque, NM) was employed to provide o
approximate size distribution of collected particulate matter at various
locations. Collection on glass substrates was used and relative density
ratios were estimated by optical comparison to determine size
relationships.
High volume particulate phase samples were acquired using several
pumping systems, all with flow rates greater that 1.3 m3/min. Samples were
collected on Pallflex Fiberfilm Type T60A20 Teflon-coated glass fiber
filters (Pallflex Corp., Putnam, CT) . Particulate concentrations were
120
-------
CQ ermined gravimetrically. The gas chromatographable major organic
^mposition of the particulate phase organics collected by filtration was
wlJjrmtned by ultrasonic extraction of the filters in toluene after spiking
c ;• an Internal standard (IS), volume reduction of the solvent, and
co y column GC and GC/MS. Two highly tumorigenic and mutagenic
f *tltuents, benzo[a]pyrene (BaP) and 1-nltropyrene (1-NPy), were isolated
norm-itW° of the ex"aust particulate extracts by semipreparative scale,
Phase high performance liquid chromatography, and were measured
caPillary column GC-mass spectrometry with selected ion monitoring
he method of internal standards.
DISCUSSION
m0l.e As expected, the diluted diesel exhaust was determined to be much
e*aBmi°nCent:rated than tne workplace atmospheres contaminated with it. For
t0 55«e' C0l»centrations of total suspended particulates (TSP) ranged upward
ana jqn'i8/'n3 in certain workplaces, although most were between 100 MS/1"3
5Q tfR/ 3 ^8/m3 , Outside ambient air particulate levels were approximately
ffom 17' *n c°ntrast, concentrations of TSP in the diluted exhaust ranged
ai2e "I'7 mg/m3 to 12 mg/m3 . Interestingly, differences in the particle
Uediatl riljutions for tlie two atmospheres were considerable. The mass
to o 5 aerodynamic diameter (MMAD) for the exhaust samples ranged from 0.2
MMAQ'J Wl ^keit with large geometric standard deviations), whereas the
with s°r-ttle workplace particulates had a distinctly bimodal distribution,
tailgln V"er particles ranging from 0.4 - 0.9 pm, and the larger particles
c°ul(j \ m 3 " * ^m< Clearly, the larger particles in the workplace
due to material from other sources, or agglomerated diesel
Phase Composition
detailed characterization of the particulate phase organic
in the one of the atmosPheres at a particular sampling
f Presented in the inventory given in Table I. The data in the
de those for both the tlme averaged (TA) samples and the time
average (TWA) calculated from the time resolved samples. The
atmosPhere was found to be a very complex mixture of both
sPectea ° and aromatic compounds. The n-alkanes were the most concentrated
entr ».« ^^ ranged from C12 to at least C33 arid were found at
n8/m3 i up to 19° ^S/8 of Tsp (corresponding to a concentration of
Vfire Jd in the workplace air) in the TA sample. Pristane and phytane also
e
tlte *lkaUtlfiedl "These two distinctive branched alkanes also were among
dl«sel Pes identified in the diesel engine exhaust (see below) and in the
' °f 8reater toxicological importance is the finding of
c0nc "^ concentrations of polynuclear aromatic hydrocarbons (PAH) .
ries nj-rat*ons of many of these PAH were as high as those of the n-
nS/mh Was 65 ^*s/5 ^13 n6/m3)' and benzo[ghl]perylene was 160 jug/g
*i coiteenl in ttle TA samPle- Most of the otner PAH measured were at least
I, °wed S,rated as BaP> The relatively high concentrations of these PAH
ter ext t0 be detected readily in the qualitative GC-MS of the crude
°UtdQor ^acts. This situation is considerably different from the ambient
°ncentra?t SlunP1e. iti which the PAH were ca. 10 -fold lower in
acion than the alkanes.
tf interesting observation was that the concentrations of the
1 * In «,Ca' °2i^ alkanea in the TR particulate samples were higher than
vhlch corresPonding TA particulate sample for one of the facilities
a« extensive set of measurements were performed. The data
121
-------
indicated that for constituents eluting before C24 , the amounts compute:
for the TWA of the TR samples were greater, but for larger (less volatile
alkanes and PAH, the differences are minimal. This suggests that so::
preferential loss of the lower alkanes occurred during the longer I
sampling-(typically 6 hours duration vs. 1 h for the TR sampling). The Pi
are more polarizeable than the nonpolar alkanes, and their losses t
sublimation from the filter media may be less than those for alkanes wit
similar boiling points and vapor pressures in their pure solid state (nc:
sorbed on particulates) , i.e., the PAH may be sorbed more strongly to thi
particulates than the alkanes. We have observed this preferential sorptk
of aromatics versus aliphatics with coal combustion stack ash. This
phenomenon illustrates the problems associated with particulate organic!
sampling by filtration, and the potential bias in data based upon long-ten
sampling periods. In some cases, however, such long sampling times art
required to provide sufficient sample for analysis.
Approximate concentrations for the identified majo:
.chromatographable particulate phase constituents for a selected exhaust
sample from an M-60 tank are presented in Table II and are calculated K
mg/g of particulate matter and /ig/m3 of diluted exhaust. The majo:
components were a series of n-alkanes ranging from C15 through at leas!
C30 at concentrations from <0.1 to 7.4 mg/g. In addition, pristam
phytane, and numerous alkylated 2- and 3-ring polycyclic aromati;
hydrocarbons (PAH) were found at much lower levels relative to the c-
alkanes. All of these compounds have been confirmed in diesel engins
exhaust* , and there is evidence5 that the hydrocarbon distribution extend!
to at least C<,0. Tentative identifications of fluorenone
dibenzothiophene, and two methyl dibenzothiophenes also were made, but ths
latter two are difficult to distinguish from C4 - and C5 -naphthalenes
(respectively) by mass spectra alone. The sum of these identified species
totaled 6.8 wt% of the particulates. Additional organic matter was preset:
but was not readily identifiable. This was indicated in the remain^
peaks and also in the baseline rise in the chromatograms generated in the
sample analyses. The latter, which was not found in the analysis of filter
blanks, probably was contributed by polar compounds which do not
chromatograph well, and by the pileup of numerous trace-level constituents
In Table III are compared the levels of nitropyrene an!
benzo(a)pyrene in the particulates of workplace air contaminated wit!
diesel exhaust and that collected from the exhaust of idling engines. I:
general, the data for the TSP levels are in good agreement with those
reported in the literature6-7, and are quite similar to each other.
Organic Vapor Phase Composition
The major vapor phase constituents determined to be present in th<
workplace atmosphere were n-alkanes from C5 through C16, benzene, and i
series of alkyl benzenes extending through CA-substitution. Naphthalem
also was detected. Not surprisingly, the workplace air sample had highei
concentrations of these constituents than the outside air samples. The M
air concentrations of several vapor phase organic compounds calculated fra
TR samples collected at a central location of one of the facilities an
listed in Table IV. Benzene was found at 5.5 to 6.0 /*g/m3 , and even highei
concentrations of toluene (36 and 49 Mg/m3) and other aromatics were
measured. The vapor phase of the idling diesel engine exhausts were
analyzed qualitatively and quantitatively as well. Compounds identified
included a range from the C6 through at least the C16 n-alkanes and include
benzene and a series of alkyl benzenes. There is some overlap ic
composition with the more volatile compounds in the particulate phase,
This overlap is probably a result of the vapor-particle partitioning of
122
-------
Compounds as wel1 as their sublimation from the filter during the
e collection. Quantitative determinations of these compounds
i| that the n-alkanes were the major compounds, although benzene and
vol *•?? a^so were relatively concentrated. The air concentrations of these
fchan v or£anic compounds were approximately an order of magnitude higher
hose of the particulate phase organics.
w°rk 1 Figure 1 are compared the gas chromatographic profiles of several
an i^i^°e a*-r samples, an outside air sample, and large diesel engines in
profn ^ fflode. There are considerable qualitative differences among the
°°tice 6S' First, the workplace atmospheres have higher proportionate
subst*1 rations °f the more volatile constituents. There are also
s&Hpl nt relative differences in the compositions of the two exhaust
VehiciS' ^e VaPor phase organics were generally more concentrated in the
ompo ? samples than in the dynamometer samples. The differences in
, can be affected by factors such as engine condition, degree of
r P' speed, and load. Comparison of the profiles does suggest that the
n P"ase composition of workplace air contaminated with diesel exhaust
e easily simulated with a large vehicle engine mounted on a test
Lon in Workplace Air Composition
j Pfltial variability in the composition of the workplace atmosphere
f ^^gated by collecting and analyzing air samples at different motor
n*0^ tie-s and also at different locations within a single facility
Same tirae Peri°d- As graphically illustrated by the gas
of tlie raaJor chromatographable particulate phase organics
fr°ii three motor pools, the particulate organic compounds were
rent quantitatively. The samples from the three facilities all
many of the same constituents, but the concentrations were
- In addition, the distribution and intensity of the unresolved
r*se during the chromatogram was quite different among the three
At the time these samples were collected, the highest
and coraPlexit:y were observed at DISCOM, followed by DOL, and
Engineering Bn motor pools. Within a given motor pool facility,
BaP a^iati°n of particle phase indicator compounds, including n-C2A
*r minor.
ComP°sition of the workplace atmosphere is not expected to be
°easfe |.v ^ rather to change with time as contaminant sources initiate and
>* ^ey e emissions, as the emissions are dispersed in the facility, and
co^ ate rem°ved by the facility ventilation. The temporal variation in
t S*t*on °f the workplace atmospheric contamination was tracked by
to t aPProxiniately hourly sequential TR samples at single locations
e*e f0 w° °f the motor pools. The samples collected at a single location
k t0 ^6 much more uniform than those from different motor pools,
expected from the greater uniformities in the ventilation and
the te and load of the work activities. These similarities are evident
tcu, gas chromatograms (Figure 3) of the major chromatographable
iate Or8anic compounds from three filter samples collected from a
10Cation in one of the facilities.
I sPh«5 ntl-tat*ve differences among the samples from both workplace
^Uato ea and vehicle exhausts are illustrated by the data presented for
t,ft replr S of atmospheric contamination listed in Table V. The indicators
V sattplfentatliVe °f the characterization studies performed during all of
oP°r Ph *nd analyses- T*16 indicator compounds include TSP, benzene (a
tu constituent), n-tetracosane (a major particulate phase
6nt which is not affected by long sampling periods) , and BaP (a
123
-------
tumorigenic particulate phase constituent) , Included in the tabulation are
data for the ambient outside air (the background) and diesel engine exhaust
from an M-60 tank (a major workplace atmospheric contamination source). It
general, the samples collected in the motor pool garages were much more
contaminated than the outside background. They also were much less
concentrated than in diesel engine exhaust. Two main differences among the
facilities were that the air concentrations of these indicators were
different by factors ranging from ca. 4 to 50, and the concentrations of
the components in the TSP also varied over an order of magnitude, i.e.,
both the air concentrations of the contaminants and the composition of the
particles was different among the three facilities. The relatively high
concentrations of BaP at DISCOM could reflect elevated diesel engine
emissions, but the n-tetracosane concentration appears lower than would be
expected from a major engine exhaust contribution. A major conclusion is
that diesel engine exhaust is not a suitable surrogate for the
toxicological study of workplace atmospheres. Other emission sources and
transformations of emissions may be important.
CONCLUSIONS
The chemical composition of workplace atmospheres contaminated will
diesel exhaust appear to be exceedingly complex. Building to building
differences occur even though the fuel source for vehicles operating in
such a facility are identical. There appear to be substantial differences
between the particle size distributions of workplace atmospheres and that
of those sources which contaminate them. Long duration sampling tends to
alter the apparent composition of the collected particle phase, and
composite samples of shorter duration may enhance compositional accuracy.
Diluted idling large vehicle engine exhaust is probably not a
compositionally accurate surrogate for workplace atmospheres for inhalation
toxicology studies.
REFERENCES
1. K. J. Bombaugh et al,, "Aerosol Characterization of Ambient Air Near
a Commercial Lurgi Coal Gasification Unit, Kosovo Region,
Yugoslavia," EPA 600/7-80-177 U. S. Environmental Agency, Research
Triangle Park, NC (1980) pp. 32-35.
2. C. E. Higgins, R. A. Jenkins, M. R. Guerin, "Organic vapor phase
composition of sidestream and environmental tobacco smoke from
cigarettes," Measurement of Toxic and Related Air Pollutants. APCA
Publisher, Pittsburgh, PA. 1987, pp. 140-145.
3. W. H. Griest, C. E. Higgins, and M. R. Guerin, "Comparative Chemical
Characterization of Shale Oil- and Petroleum-Derived Diesel Fuels,*
in Proceedings of the Twenty-Forth Hanford Life Sciences Symposium.
Pacific Northwest Laboratory, Richland, WA, in press.
4. F. W. Karasek, R. J. Smythe, and R. J. Laub, J. Chrom. 101. 125
(1974).
5. Black and L. High, "Methodology for Determining Particulate and
Gaseous Diesel Hydrocarbon Emissions," SAE Automot. Eng. Tech. Pap,
Ser. 790422 (1979).
6. R. L. Williams and D. P. Chock, Environ. Int. 5_, 199 (1981).
7. T. L. Gibson, Atmos. Environ. 16. 2037 (1982).
124
-------
Table I, Inventory of ParticuLate Organic Compounds Identified
in the Workplace Atmosphere in DI3COM in September 24, 1986
Jjt.ntl
*
on
min-
.
SO, 7
.
52,
m/z
198
Concentration0
IA
Identification
ng/m
TWA of TR
/fc/g ng/m
n-Dodecane
Diethylphthalate
n-Hexadecan*
Tributylphosphata
n-Heptadecane
Priatane
Fhenanthrene
n-Octadecane
Phytane
C2-Acenaphthene
n-Nonadecane
Dibutylphthalate
n-Eicosane
C2-Fluorene
C5-Naphthalene
Fluoranthene
Acephenanthrylene
Pyrene
n-Heneicocane
C2-Phenanthrene
n-Doco»ane
n-Tricosane
Dibutylbenzylphthalate
Benzo(ghi)£luoranthene
n-Tetracoaane
C^~Phenanthrene
Eenz(a)anthracene
Chryaened
C1-Bento(ghi)fluoranthene
C2-Pyrene
n-Pentacoaane
2,2'-MethyLenebia(4-ethyi,
6-t-butylphenoL)*
Octylphthalate
C1-Chryaene
Cyclopantachryaane
n-Hexacoaane
n-Heptacoaane
Phthalate
Beneo(b/J)£luoranthenea
Benzo(k)fluoranthened
n-Octacoaane
Banzo(a)fl,uoranthene
Benzo(e)pyrene
Benco(a)pyrene
Perylene
3.8
23
14
24
-
29
-
51
150
-
190
130
120
-
68
100
84
88
62
.
140
28
42
24
48
65
"PParBnt molecular ion.
0.8
4.9
-
2.8
4.7
5.7
10
30
37
26
23
13
20
17
17
12
.
27
6
8.3
S
9
13
17
62
19
57
59
140
150
291
300
190
110
59
107
70
52
61
.
73
18
29
25
57
83
16
4.7
14
15
34
38
77
77
50
28
16
26
19
14
17
-
17
4
7.7
6
14
20
"N8S4"l«ll v ln DI8-2*-TA-3 (0855-1559 hra) and TWA of DIS-24-TR-1 through -TR-5
•T *tttiti,d j B) «««pt for PAH (-TR-2 miaaing).
*at«tiv, tin *BP«*«ta OC-MS analyaia apacific for PAH.
"•ntifioation from apectraL matching.
125
-------
Table I. Inventory of Patticulate Organic Compounds Identified
in the Workplace Atmosphere in DISCOH in September 24, 1986
(Cont'd)
Concentration0
TA
TWA o£ TK
Retention
Time, min.
62.
63.
64.
64.
65.
65.
67.
-
67.
68,
66.
69.
69.
70.
72.
74.
a
5
Z
4
1
9
4
8
3
7
2
3
ft
e
7
m/za
57
266
254
264
57
191
57
276
276
276
278
57
276
276
288
Identification53
n-Nonacosane
Cj-Benzopyrene
C2-Benzo(ghi )f luorantbenie
A Iky 1- PAH
n-Triacontana
Heterocyclic
r, -Henetriacoritane
Dib anz( a. j ) anthracene
IndanoClZ3-cd)pyrene
Dibenz(a, c / a, h) anthracenes
PAH
n-Dotciacontane
Eenao(ghi Jperylene
Anthanthrene (7)
n-Tritriacontane
Alfcyl-PAH
Pe/s
99
-
-
-
50
-
130
16
70
5
_
47
160
-
44
"•
ng/m3
20
-
-
-
10
-
25
3
14
1
_
9.2
32
-
8.6
-
(JKfe.
75
-
-
33
_
84
11
140
10
_
34
120
35
ng/o'
21
.
10
23
3
34
I
9,1
25
9.!
Table III. Comparison of Benzo(a)Pyrene (BaP) and l-Nitropyrene (1-NPy)
in Workplace Air and Diesel Exhaust Samples
Sample Type
Ambient Air-Background
Motor Pool Workplace Air
Dynamometer Exhaust (M-60 Tank Engine)
Vehicle Exhaust (M-60 Tank, Vehicle)
BaP
TSPa ng/m3
2.4
6
11
17
0.30
3.5
63
48
micrograms per g of total suspended particulates.
1-NPy
TSPa ng/i:
<0.4
4.5
0.32
2.1
1,0
u
5.7
126
-------
Table II
'"^quantitative Determination of Major Pacticulate Phase Organic Compounds in Exhaust
of H-60 Tank (Sample 25-A-l)
No.
1
2
3
4
S
6
7
S
a
10
11
12
13
14
IS
IB
17
18
la
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
33
36
37
38
39
40
41
+2
43
44
4S
46
+7
SO
Concentration
Identification'
mg/s
Cg-Naphthalene
n"C15 H32
Fluorene
n"c!6 B34
Hydrocarbon
Hydrocarbon (Maybe 2-Methyl C^g)
Hydrocarbon
Hydrocarbon
C2 -Ac enaphthalene/Cj -Fluorene
C2-Acenaphthalen«/Cj -Fluorene
n-C17 H3e
Pristane
Fluorenone
Dibenzothiophene (C^-Naphthalene? )
Hydrocarbon
Hydrocarbon (Maybe 3-Methyl C17)
Phenanthrene
Anthracene + Hydrocarbon
n-C18 H38
Phytane
C^-Dlbenzothiophene (C^-Haphthalene? )
Hydrocarbon
C^-Dibanzothiophene (C^-Naphthalene? )
C^ -Phenanthrene
n"clfl H40
2-Methyl Phenanthrene
Ce-Haphthalane
Hydrocarbon
n"c20 "*2
C o — Ph en anth r en e
C2-Phenanthrene
C2-Phenantbrene
Fluor an then*
Cy-Naphthalena
Hydrocarbon
Cj'Phenanthrene
n-C21 E\4
Pyrene
Cg-Fhananthrene
Cg-Fhenanthrene
n"C22 H46
B enzo (fa ) f luor en e
n"C23 H46
__« H
C24 H50
n"C25 H52
n"C26 H54
n"C27 H58
n"C28 H58
n"c29 H60
n"C30 H62
< 0.1
0.17
< 0.1
1.0
0.2
0.2
- o.i
' 0.1
< 0.1
< 0.1
5.3
1.2
< 0.1
" 0.5
" 0.2
0.1
1.5
< 0.1
4.5
1.0
- 1
" 0.4
- 0.7
" 0.7
6.9
2.2
- 0.7
- 0.7
7.4
- 2
" 2
- 0.9
- 0.6
" 0.8
" 0.3
" 0.3
7.0
" 1.1
3
1
4.8
0.5
3.1
1.5
1.0
0,35
0.15
< 0.1
< 0.1
< 0.1
< 0.3
0.48
< 0.3
2.6
0.56
0.56
" 0.3
" 0.3
« 0.3
« 0.3
15
3.4
*• 0.3
- 1
- 0,6
0.3
4.2
< 0,3
13
2.8
" 3
- 1
" 2
" 2
19
6.2
" 2
- 2
21
- 6
" 6
" 3
" 2
- 6
' 0.8
" 0.8
20
" 3.1
8
3
14
1
8.7
4.2
2.8
0.98
0.42
< 0.3
< 0.3
< 0.3
^•nt tiliciitoris *r* tentative and other ikomera are possible.
***°luti i0n *>tiraat*« should be considered seni_1>ioaui* of tn«ie higher relative concentrations. Unita are mg per g of
matter and pa per n of diluted exhaust.
127
-------
Table IV. Time Weighted Averages of Vapor Phase Organic Compound*
in Time Resolved Samples Collected in the Workplace Atmosphere
at DISCOM on September 23 and 24, 1986
Compound
Hexane
Benzene
Heptane
Toluene
Octane
Ethyl Benzene
m/p-Xylenes
"0907-1242 hrs calculated
b 1120- 1611 hrs calculated
Concentration, A*g/m3
AM, 9/23/86a PM,
13
5.5
11
49
6.9
19
220
from DIS-23-TR-1 through -TR-3.
from DIS-24-TR-3 through -TR-5.
9/24/8'
13
6.0
9.9
36
4.1,
9.3
71
Table V. Comparison of Contamination Indicators for Three Motor Pool
Workplace Atmospheres, Outside Ambient Air, and Diesel Engine Exhaust
Indicator Concentration0
ISP,
Location Mg/n>
DISCOMb 270
DDL* 110
4th Engineers 46
Outside Background8 46
Diesel Engine Exhaust*1 2,600
n-Tetraoosane BaP
Benzene,
(*g/m ng/m3 pg/g ng/m3
5.5C 28 110 20d
HA 7.4 67 3.2
HA 12 230 0.4
NA 3.4 71 0.4
220 4,200 1,500 48
*/.
83d
29
7.1
8.0
17
*NA - not analyzed
bTWA of DIS-24-TR-1 through -TR-5 (0584-1611), except «» noted.
'TWA of DIS-24-TR-3 through -TR-5 (1120-1611).
Same as B, except -TR-2 was missing.
*DOL-25-TA-3 (0653-1531).
fTHA of EHO-30-TR-1 through -TR-5 (0859-1536).
'DOL-Outiide Background, 9/25/86.
"M-60 tank, 25-A-l,
128
-------
Comparison of the Major Vapor Phase Organic Compounds in
Ambient Outside Air, Workplace Air at the 4/68th Armored Div
and Forth Engineering Bn Motor Pools, M-60 Tank and exhaust
from an M-60 Engine on a Dynamometer Test Stand.
129
-------
4th ENGINEERS
Figure 2
20
40 60
TIME (min)
80
100
Comparison of the Chromatographable Major Organic Partictil*
Phase Organics from the Workplace Atmospheres at the Forth
Engineering Bn, DISCOM, and DDL Motor Pools. x
Figure 3
1353-1611
60
TIME (min)
.ytf
Comparison of the Chromatographable Major Organic ParticulflV
Phase Organics in Time Resolved Samples at a Single
in the DISCOM Motor Pool.
130
-------
•MOTHCAKE STUDIES
AIR QUALITY TEST HOUSE
re'c' °layton and E- EuQene Stephenson, Jr.
Res&archTrlanglenpark,NC 27709
-f ^ackson and Leslie E. Sparks
Air and p mental Protection Agency
!ray Engineering Research Laboratory
Triangle Park, NC 27711
p H,has leased a tyPica' three-bedroom, two-bath, one-story frame house in the Research
of NC> area for incloor air quality OAQ) research. The test house Is used to evaluate the
resident!af m ns *rom common household and building materials, to determine the effects of
ac
Va"date uJ^'t'es which release organic emissions, to evaluate IAQ control technologies, and to
used to D H mode'Si Tne results of testing in EPA's laboratory-size environmental test chambers were
^othcak P°"utant concentrations in a residence through a model developed by EPA. This
e study was the first conducted in the test house to verify the model.
Mothca|Ihe house was sampled for levels of 1,4-dichlorobenzene using GC/ECD and GC/FID.
cake per ioW^re P'aced in a bedroom closet according to the manufacturer's recommendations (one
^droom H °f closet volume)- Syringe samples were taken at four locations in the house (closet,
^C/ECD a ^acent bedroom, den) and one outside location. Samples were injected directly, for
1(16 closet9"3^8'8'and orrto Tenax sorbent for later analysis with GC/FID. CO2 was also injected Into
t6rnPeratud measured by an NDIR monitor to characterize dispersion patterns in the house. Inside
Wind dirertf and nurnidity data were recorded, as well as outside meteorological data (wind speed,
ection, temperature, humidity, barometric pressure, and solar radiation).
Predjo hlii8 mothcake study demonstrated that small chamber data can be used with a model to
Pollutant concentrations in the test house.
131
-------
Introduction
As AEERL on-site operations contractor, Acurex is involved with EPA's Indoor Air Program^
particularly in the small chamber laboratory and the test house studies. This paper describes the in**
mothcake work performed in the test house and how it fits into the overall program.
This mothcake test was the first made in the test house which could be used to verify a
developed by EPA to predict indoor concentrations of pollutants based on small chamber emission
factors.
Initial mothcake work was done in 1986 in the small chamber laboratory at EPA's
Environmental Research Center. Four-ounce cakes of 99.75% pure para-dichlorobenzene were
purchased from a local department store. The results indicated that the emission rate was very
dependent on the temperature and air exchange rate, but relative humidity showed little to no effe*
addition, the emission rates were essentially constant at fixed temperature and humidity. The j
experimental techniques are defined by Nelms.1 The emission factor determined for the condition**
the test house experiment was 1 .4 mg/cm2-h.
u40
A model was developed by EPA.2 to translate the results gathered in the small chamber w^
into predictions of concentrations in a residence. This model estimates the effects of heating,
ventilating, air conditioning, air cleaning, room to room air movement, and natural ventilation on
pollutant concentrations.
The model began with a single-room mass flow described by:
Amount In
Accumulated
Produced Removed
Amount Out
Sources of pollutants may be in the room, in the HVAC system, or outside the room. Pol
may be removed by sinks, which may be in the room or HVAC system and may become sources
the pollutant concentration drops below a given value.
For this single room, the pollutant concentration is determined by a mass balance of the
pollutant flows:
Amount Accumulated = (Amount In) + (Amount Produced) - (Amount Removed) - (Amount Out)
This analysis can then be expanded to multiple rooms In a house by a system of equation^
each room. Assuming no pressure buildup in any room, the amount of air entering each room thrO0*
all sources (HVAC, outside, and other rooms) must equal the amount of air leaving the room.
The type of mixing must be specified for the model, but the exact mixing cannot be specif*0
because of its complexity; therefore, some assumptions must be made. Two common mixing
simplifications are described by the plug-flow and well-mixed models.
Plug-flow assumes a pollutant concentration gradient along the airflow path, while the
mixed model assumes a constant pollutant concentration throughout the room. Because previous
test house studies showed little variability of concentrations in a room, the well-mixed model was
selected for this effort. Now we have the small chamber data and the model, but need to close tri*
through model verification.
The AEERL test house was established to provide a full-scale test facility for IAQ studies*
fully described by Jackson, et a!., in "EPA's Indoor Air Quality Test House, 1 . Baseline Studies."3
This characterization work in the test house had been conducted during the kerosene
testing, including blower door and SFe tracer gas studies for air exchange rate; migration testing
C02 to determine the pollutant movement between the comer bedroom and the den; and strati "
testing to determine any vertical gradients of pollutants in the room. This was accomplished by
sampling elevations comparable to lying on the floor, sitting, standing, and at the ceiling level.
132
-------
0-35 Acw'f excnan9e rate testing by SF6 tracer gas method indicated a nominal infiltration rate of
H (air changes per hour) for the weather conditions during the mothcake study.
Tfist Description
Tnjs sh i '3lan ca"ed '°r sampling to begin 3 days after placement of the mothcakes in the closet.
on dav °!i allow su'ficient time to achieve steady state conditions in the house. Sampling would occur
Amovedi« 6> 8l and 11 after test starti Immediately after sampling on day 11 the mothcakes would be
and is T0m the closet and tne concentration degradation monitored by sampling on days 12,14,16,
*>• The house would be aired out for 3 days prior to the start of subsequent tests.
rnanufa J88tif19 began by placing the mothcakes in the comer bedroom closet, according to the
taken o t i*fS specifications (one mothcake per 12 ft3 Of closet volume). Meteorological data were
the HVAr tlle nouse to characterize weather effects. Inside temperature was maintained at 26°C by
Ian rem i'and relat've humidity was uncontrolled but recorded by a meteorograph. The HVAC system
mained on throughout the test period,
and on ee days after Placin9
-------
Initial and final weights of the five mothcakes showed that the amount emitted during the 11
days of testing was 204.15 g. The beginning surface area exposed was 570 cm2. An average
p-dichlorobenzene emission factor was calculated by dividing the total weight lost of the mothcakes by
the duration of the experiment. The average emission factor was 820 mg/h or (1.44 mg/cmZ-h), which
is in excellent agreement with the small chamber emission factor.
The results of the direct injection sampling are shown in Table I.
The differences between day-to-day measurements with mothcakes in place in Table I are
probably normal data scatter; therefore nothing should be inferred from the apparent trends. Since
concentrations were still -0.5 mg/m3 in the house 7 days after the mothcakes were removed, the data
suggest that a significant sink exists for p-dichlorobenzene in the test house. If no sink were present,
the levels would not be detectable due to dilution or infiltration.
Analytical data from the Tenax tubes were less consistent than the direct injection results.
These results are shown in Table II.
The CO2 concentration data were evaluated and found to be fairly consistent from day to day
during the testing. This indicated that, for the 7 days that mothcakes were in the house, there were no
significant changes in infiltration rate. The meteorological data support that assumption.
Now. if we return to the model predictions, we find that the emission rate, E(t), is defined by
E(t)= dc/dt + Qct &
where dc/dt = change in p-dichlorobenzene concentration as a
function of time
Q= airflow through house
ct = concentration in house at time t
An effective emission rate was obtained by smoothing the concentration data, then calculating
dc/dt from the smooth curve. The effective emission rate was then calculated and an average value
obtained The average emission rate was 480 mg/h compared to 820 mg/h based on the weight loss of
the mothcakes. The difference (340 mg/h) between the emission rate calculated by weight loss and the
effective emission rate calculated from the house concentration data is the sink term.
Initial model runs with the sink term and estimates of the room-to-room airflows were in good
agreement with the measurements for all rooms except the closet. The closet concentrations were a
factor of 13 too high (see Case 1, Table IV). Also, the den and master bedroom concentrations were
too low relative to the corner bedroom concentration. Additional model runs showed that the flow from
the closet to the corner bedroom was the key unknown parameter for determining the p-
dichlorobenzene concentrations in the house.
Therefore, experiments were conducted to define the airflows in the test house and to estimate
the type of mixing. During these experiments, the air-handling system flows were measured. Flow
visualization studies to determine the nature of the in-room and room-to-room mixing were conducted
with neutral density balloons and with neutral density helium bubbles.
The measured flows of the air-handling system were found to range from 38 m3/h in the middle
bedroom to approx 280 m3/h in both the corner and master bedrooms, and 679 m3/h into the den.
The balloon and bubble experiments showed that, even with the air-handling system on,
considerable mixing existed between rooms, it was expected that all of the airflow would go from the
corner bedroom directly into the return vent located in the hall, but some flow directly into the master
bedroom was observed. These experiments also indicated that there was a substantial airflow into and
out of the closet. Finally, the visualization studies indicated that there was a flow between the closet
and the hallway, and between the closet and the master bedroom.
134
-------
Hot-wire anemometer measurements were made of the airflow velocities through the cracks
n and under the closet doors. These measurements showed that the airflow into and out of the
was between 4 and 9 m3/h.
inPut d f^°del calculations were performed with the input data from the experimental studies. The
hours ft! i °r tfle model mn are shown in Table III. The model calculations were stopped after several
ot simulated time because a steady-state was reached.
and tho results of this run are shown as Case 4 in Table IV. The agreement between the model
ne measured data is excellent for all rooms.
tyerQ Several additional runs were made to determine the effects of errors in the input data. Runs
flows u With low flow from the closet-no sink-a Iar9e sink.and errors in estimating the inter-room
Model results are summarized in Table IV.
Delusions
rrtodQi mothcake study described here demonstrated that small chamber data can be used with a
betwee S.redict P°llutant concentrations in the EPA test house. Excellent agreement was found
dorninat Sma"cnamber and test nouse emission factors. The study also showed that, when IAQ is
various i by a 'arQe ^'^ source> knowledge of the source strength and rough estimates of the
mucr, of (|["0ws are sufficient to predict concentrations within reasonable (± a factor of 2) accuracy for
rooms Building. The most important airflow is that from the room with the point source to adjacent
Finally, the study showed that, when data are available for most parameters important for the
6 model predictions are in excellent agreement with the measured values.
Possiblei* model is a powerful tool for evaluating IAQ control options. The ease of use makes it
9 to run the model several times to determine effectiveness of control options.
'actors onal research is necessary to prove the feasibility of using small chamber emission
'^Portant tfle model to predict IAQ for complicated situations. Additional work on sinks is especially
The model will be refined as more data become available.
I" H. Nelms, M. A. Mason, B. A. Tichenor, "Determination of emission rates and concentration
levels of p-dichlorobenzene from moth repellant." Presented at 80th Annual Meeting of APCA,
New York, NY. (1987).
2.
L-E- Sparks, M. D. Jackson, B. A. Tichenor, "Comparison of EPA test house data with
Predictions of an indoor air quality model." Presented at ASHRAE-IAQ 88, Atlanta, GA. (1988).
3.
M. D. Jackson, R. K. Clayton, E. E. Stephenson, W. T. Guyton, J. E. Bunch, "EPA's indoor air
Duality test house, 1. Baseline studies." In Proceedings of the 1987 EPA/APCA Symposium on
Measurement of Toxic and Related Air Pollutants. Research Triangle Park, NC. (May 1987).
j=: M. Hansen, "Protocol for the collection and analysis of volatile POHC's using VOST." EPA-
600/8-87-007 (NTIS PB84-170042). (1984).
135
-------
TABLE I. RESULTS OF DIRECT INJECTION SAMPLES (mg para-dichlorobenzene/m3 a&)
Day
Closet
Corner Bedroom
Room
Master Bedroom
Den
Baseline
4
6
8
11
Mothcakes Removed
12
14
16
18
107
53.6
70.9
53
5.4
1.5
1.3
1.0
4.72
4.41
5.51
5.61
2.11
0.90
0.52
0.42
3.49
3.50
4.18
4.27
2.07
0.77
0.65
0.55
3.84
3.30
3.80
4.02
1.80
*
0.45
0.45
_^,&
* Sample tost
TABLE II. RESULTS OF TENAX INJECTION SAMPLES (mg para-dichlorobenzene/m3^
Day
Closet
Corner Bedroom
Master Bedroom
Dfll>
Baseline
4
6
8
11
Mothcakes Removed
12
14
16
18
0
82.3
57.4
34.2
63.1
0.69
1.55
1.03
0.29
0
5.42
4.51
3.96
5.02
1.81
0.79
0.39
0.21
0 <
3.47 3.
2.94 I'1
1.82 3.!
4.17 3-I
1.95 OJ
0.77 &|
0.34 &
0.21 to
^
TABLE III. INPUT DATA FOR MOTH CRYSTAL ANALYSIS
Source strength 1.4
Air exchange with outside 0.35 ACH
Air exchange between closet and bedroom 4 m3/h
Air-handling system airflows defined above
All airflow to air-handling system is from hallway
Air exchange with outside is evenly divided between rooms
Sink removes 40% of material
TABLE IV. SUMMARY COMPARISON OF MODEL PREDICTIONS AND
Case
Closet
Ratio of predicted to measured concentrations
Corner Bedroom
Master Bedroom
1
2
3
4
13
1.1
3
0.998
0.96
0.048
1.8
0.89
0.85
0.025
2.2
0.97
Measured concentration Is average of alJ measurements
Case 1: Low flow from closet and sink (initial run). Case 3: 4 m3/h flow from closet
Case 2: Low flow from closet and large sink. Case 4: Measured flows and sink.
136
-------
1 GOALS
RESULTS OF THE BALTIMORE
STUDY:
AND STUDY DESIGN
Wi 1 1
iaitl C* Nelson, Andrew E. Bond and
n
U. s "meital Monitoring Systems Laboratory
Hese' Environmental Protection Agency
ch Triangle Park, NG 27711
.. Manale and Lance A. Wallace
shi Jnviron"iental Protection Agency
Lt>gton D l
on, DC
rpu
, Baltimore TEAM VOC Study was a cooperative effort involving
a ® Risk Identification Division (RID) of the Office of Policy, Plan-
Evaluation (OPPE) as part of its Integrated Environmental Manage-
(IEMpK Region III, and the State of Maryland, City of
^altimore County and Anne Arundel County. The Baltimore area
because it has a significant air toxics problem, but
tne TEAM approach in a metropolitan area not having major
o Vi
Nei L°a"'~ industry facilities, as did previously studied California
ersey areas.
t • The
loti ej. "^In objectives of the Baltimore Study were to estimate popula-
e*Posur cSure to selected VOC's, compare indoor, outdoor, and personal
IfiMp " an<^ ^° coordinate the exposure monitoring data collection with
x ^iqn '*"e^ study. The study design utilized probability sampling
Ot>*ioo
-------
Introduction
Previous Total Exposure Assessment Methodology (TEAM) Studies
been performed during 1980-8^ in several locations including Bayonne
Elizabeth, NJ and Los Angeles, Antioch, and Pittsburg, CA.1>2>3
communities have often been areas of industrial or chemical manufacturing
Pollutants measured have been approximately 20 target volatile org
compounds (VOC's), In all, more than 700 participants have provided
measurements. The major goal of the study has been to develop and
chemical and statistical methodology for estimating human exposure
selected toxic or hazardous substances. The probability sampling appro
used to select study participants permits inferences for the communi#
residents, not Just the study subjects. Direct measurements of the
were made on the air each person breathes , outdoor air in the
and the water available to drink. Concurrently, the same chemicals
measured in each person's breath.
Previous studies have indicated that personal and indoor
to many VOC's exceeded outdoor concentrations, breath measurements
many VOC's were significantly correlated with preceding air exposures,
personal activities such as smoking, use of room air deodorizers, weal"
dry-cleaned clothing, and use of hot water were related to
exposure for benzene, para-dichlorobenzene, tetrachloroethylene, and
roform, respectively. ^>5
Methods
Baltimore, Maryland was selected as a location to extend the
Study results because it was representative of many large urban areas *.;
it also provided an opportunity to coordinate data collection with an °K
going EPA study. EPA had been conducting an Integrated Environment^
Management Project (IEMP) since 1983 with State and local governments ^
the Baltimore region. The overall purpose of the IEMP is to identify *%
assess the significance of environmental issues of local concern, to 6 •:.
priorities for acting on those issues, and to analyze appropriate &H
proaches for managing the problems that are identified. The IEMP utiliz
a multimedia approach to measure the complex, interactive effects
pollutants in air, water, and on land; and to identify effective,
ical methods of controlling those pollutants,
The Baltimore area communities selected for the TEAM Study - ^
Dundalk, located near the southeast corner of Baltimore City, and *;
neighboring towns of Parkville and Overlea, located adjacent to *J
northeast corner of Baltimore City, and considered to be a relatively M
exposure area, Dundalk, approximately eight miles south of Parkvil*^
Overlea, is downwind of an industrialized area. Thus, these study a*6 j'
were expected to be representative of a wide range of exposure conditi"
within the Baltimore metropolitan area.
The probability sampling scheme for this study was a stratif* (
multistage design in which the final sample of individuals cons
nearly self-weighting random sample from each of the study areas.
first stage sampling units were census block groups. These units
stratified into seven geographic strata and two socioeconoraic strata O
geographic stratum. Twenty first-stage units in each area (UO tot**|
were selected after sorting by size, A cluster of approximately J
household units was then selected for screening within each first-s**^
unit. Approximately six individuals per first-stage unit were seie°
138
-------
6 e"^ a^ the time of screening and asked to participate in the
e objective of obtaining approximately 160 participants (80 in
^ne two areas; approximately four from each of the ko strata) was
ari(1 ?Qd" Talale 1 shows the number of household and individual contacts
DJVS-, 4 e resulting response rates, which were relatively consistent with
r TEAM Studies.
the six-week period from mid-March through April, 1987, 155
in theiVS of the Dundalk and Parkville/Overlea communities participated
C0ria ™EAM Study. One participant from each home was selected for two
The , utlve 12-hour personal air samples and for three breath samples.
21^ rea/th samples were obtained at the beginning, middle, and end of the
Vere r Personal air sampling period. The personal air and breath samples
sPect c^ec* on Tenax and analyzed for VOC's by gas chromatography/mass
c°lle i"16^1^* In addition, two indoor and two outdoor air samples were
et* a^ residence, using SUMMA polished canisters. A selected
e
°^ c°llocated Tenax samples were collected in conjunction with the
f canisters for method comparison. This study was also augmented by
Xe<* s^ roof-top canister air samplers, one in each neighborhood,
Thea £°Hected 2U-hour samples for each day of the six-week study.
wo stations were included in order to provide appropriate ambient
PPedi t allov the IEMP program to evaluate the effectiveness of its
ive dispersion models for air quality for the study time period.
atia]_y ^ wa^er samples were collected from a small subset of homes and
tor, ^T ^or VOC's by a gas chromatograph with a Hall conductivity detec-
ne water samples were analyzed by the EPA Region III Chemical
Q_?ry in Annapolis, Maryland. The air exchange rate of each of the
es was also determined during the sampling periods using perfluor-
(PFT) compounds.
, ble 2 summarizes the chemical monitoring design and shows the
Dumber of samples collected. Including quality control and quality
measures, more than 700 canister samples and 800 Tenax samples
cted and analyzed. The summary of target chemicals that were
in the personal and breath samples ia shown in Table 3« A
group vas analysed for the indoor and outdoor canister samples
™e addition of vinyl chloride, vinylidene chloride, and methylene
th an(i the deleti°n of ^-pinene, 1,2-dichloroethane, and l,U-dioxane,
v ^inking water samples, only the following eight compounds were
n^?: chl°roforni, bromoform, broraodichloromethane, tetrachloroethyl-
dibromochloromethane, chloroethane, and 1,1,1-tri-
^-^^B-ry results of the outdoor and indoor canister data and of
and Pers°ial exposure data are discussed separately. °»7 Final
are not yet available.
od }imlteci number of formaldehyde samplers utilizing the EPA-TeJada
fl tDlJPHv/ Silica gel) were also deployed in the Baltimore Study area.
J!ed 3ite outdoor formaldehyde levels were consistently low (below 2
sites for every day measured. A small number of indoor and
outdoor (backyard) samples were also collected which indicated
in
-------
levels were consistent with concentrations measured by Stock in HousW'
Texas. ^
The preliminary results of the drinking water analyses revealed
only three of the eight chemicals were measured at levels above t
quantifiable limit. The three VOC's were chloroform with a mean level0:
23.7 ug/L (range: 17.3 - 35.2); broiaodichloroiaethane with a mean °*
9-3 ng/L (range: 6.7 - 13.^); and dibromochlorowethane with a mean "
2.7 |ig/L (range: 2.3 - 3.5). The relatively small range of th^e
values was consistent with our belief that the drinking water exposur8'
since it originates from a central supply source, should be reasona^
uniform for all study participants. The three compounds measured ^
,also consistent with previous TEAM Study results in other locationS<
Conclusions
A major TEAM VOC Study has been recently conducted in
Preliminary results indicate a general consistency with previous finding8'
Table I. SCREENING AND INTERVIEWING RESULTS
Activity.
Households Eligible
Screening Completed
Individuals Eligible
HH Questions Completed
Monitoring Completed
Dundalk
266
253
118
81
77
Parkville-
Overlea
328
295
132
81
78
Total
59U
548
250
162
155
percejj
100.0
92.3
100.0
6h>"
62.0
>
Table II. VOC MONITORING PLAN
Location
Personal
Breath
Indoor
Water
Out door /Backyard
Outdoor/Fixed Site
Method
Tenax
Bag/Tenax
Canister
Bottle
Canister
Canister
Frequency JJui
2 12-hour '.
3 samples 1
2 12-hour -
1 sample
2 12-hour -
1 2U-hour
1*65
310
10
140
-------
References
1* Wallace, L.A., The Total Exposure Assessment Methodology (TEAM)
Study, Volume I, Summary and Analysis, EPA 600/6-8j-002a, 198],
2. Wallace, L.A. , The Total Exposure Assessment Methodology (TEAM)
Study, Volume II, Elizabeth and Bayonne, New Jersey, Devils Lake,
North Dakota and Greensboro, Horth Carolina, EPA 600/6-87-Q02b
1987.
3. Wallace, L.A,, The Total Exposure Assessment Methodology (TEAM)
Study, Volume III, Selected Communities in Northern and Southern
California, EPA 600/6-87-002c, 1987.
k. Hartwell, T.D., E.D. Pelliszari, R.L. Perritt, R.W. Whitmore, H.S.
Zelon, L.S. Sheldon, C.M. Sparacino, and L. Wallace, Results from
the TEAM Study in Selected Communities in Northern and Southern
California, Atmospheric Environment, V. 21, pp. 1995-20Q1*, 198].
5. Wallace, L.A. , E.D. Pellizsari, T.D. Hartwell, R.W. Whitmore, C*
Sparacino, and H. Zelon, Total Exposure Assessment Methodology (TEAM)
Study: Personal Exposures, Indoor-Outdoor Relationships, and Breath
Levels of Volatile Organic Compounds in Hew Jersey, Environment
International, V. 12, pp. 369-387, 1986.
6. Wallace, L.A., A. Manale, and W.C. Nelson, Preliminary Results for
the Baltimore TEAM Study: II. Personal Air and Breath Measurements,
APCA Proceedings this volume (1988).
7. Manale, A., W.C. Nelson, and L.A. Wallace, Preliminary Results for
the Baltimore TEAM Study: III Indoor and Outdoor Canisters, APCA
Proceedings this volume (1988).
8. Stock, T.H., Formaldehyde Concentrations Inside Conventional Housing
JAPCA, V.37, PP. 913-918, 1987.
142
-------
PRELIMINARY RESULTS OF THE BALTIMORE TEAM STUDY
III. INDOOR AND OUTDOOR CANISTER MEASUREMENTS
A. Manale9, L. Wallaceb, and W. Nelsonc
a Office of Program Planning and Evaluation
Washington, DC
b Office of Research and Development
Washington, DC
c Environmental Monitoring Systems Laboratory
Office of Research and Development
Research Triangle Park, NC
RESULTS
For a subset of the 26 volatile organic compounds, we
estimated the percent detected for each twelve-hour sampling
period indoors and outdoors at both Dundalk and Parkville-
Overlea (Table I). In general, there was little (i.e., less than
10%) day-night difference. However, some chemicals (particularly
chloroform, trichloroethylene, and para-dichlorobenzene) were
detected much more often indoors than outdoors.
Dundalk. For six of twelve compounds, mean concentrations
indoors exceeded outdoor mean concentrations by a factor of two
or greater (Table Ila). The highest arithmetic mean
concentrations (34 ug/m3 night and 78 ug/m3 day) were found for
vinylidene chloride for both the nighttime and daytime 12-hour
periods. However, a coeluting chemical could be responsible for
some or most of the observed concentration. (19 samples analyzed
by GC-MS confirmed the presence of vinylidene chloride, but only
at concentrations of < 1 ug/m3.) For most of the subset of
coipounds presented in Table Ila, there were only slight day-
night differences in mean concentrations. However, vinylidene
chloride and, to a lesser extent, m-xylene and ethyl benzene
143
-------
displayed significantly greater mean concentrations in th
daytime samples than in the nightime samples. p-Dichlorobenzen
and methyl chloroform displayed higher mean concentrations a
night.
For outdoor samples, there were few day-night differences!
arithmetic mean concentrations. The highest mean values were fo
vinylidene chloride [Please note the caveat above.] The range o.
values tended to be less extreme for the outdoor samples than fcj
the indoor samples. Outdoor mean concentrations at the subjects
homes fell within a factor of about two of the raeai
concentrations at the two fixed sites.
Parkville. Indoor concentrations exceeded outdoj
concentrations by a factor of two or greater for nine of
twelve compounds (Table lib). Chloroform exhibited the greates
difference in indoor versus outdoor levels (9.6 to 0.2 »g/i
night and 7.5 to 1.3 ug/m3 day). Only p-dichlorobenzene, ethyl*
benzene, and 1,1,1-trichloroethylene displayed greater than
factor of two difference in nighttime mean values compared t
daytime mean values.
As in Dundalk, the range of outdoor values tended to h
narrower than for the indoor samples. Also as in Dundall
outdoor mean concentrations at the homes fell within a factor o
two of the values at the fixed site.
correlations. In Table III, we present estimates of th
correlation between outdoor (backyard) concentrations with tt
corresponding fixed site values for six compounds. We find tb,
strongest correlation (0.7) for benzene in Parkvi lie-Over lea,
The difference in the locations of the fixed sites relative faj
the backyards may account for the better correlations d
Parkville-Overlea than in Dundalk. In Parkville-Overlea, tk
fixed site was centrally located, whereas in Dundalk the fixei
site was located at the edge of and upwind of most of the
backyard monitoring sites. These results suggest that
significant daily variation in ambient levels can exist evet
within neighborhoods.
Table IV shows the correlation between the ambient concen-
trations of ten compounds at Dundalk and Parkville-Overlea. Verj
little correlation was seen for vinylidene chloride. Sons
(greater than 0.3) temporal correlation exists for benzene, ethyl
benzene, and tetrachloroethylene. Strong correlations would
suggest that one or more sources of emissions affect both areas.
We also correlated the total concentrations of all twenty-
six compounds examined in Dundalk with those from Parkville-
Overlea. We find good correlations both outdoors (0.8 at night,
0.7 during the day) and indoors (0.85 at night, 0.63 during th
day) . This suggests that the key sources of exposure are similar
for the two areas.
144
-------
TABLE Ila
Summary Statistics for Selected Compounds
by Monitoring Area, Location, and Twelve-Hour Period
in ug/m3
DUNDALK
Compound
Benzene
Carbon Tetra-
chloride
Chlorobenzene
Chloroform
o-Dichlorobenzene
p-Dichlorobenzene
Ethylbenzene
Methyl chloroform
Tetrachloroethylene
Trichloroethylene
Vinylidene
chloridea
m-Xylene
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Indoors
Nioht
0.4-27.4
4.7
0.03-8.9
0.3
0.4-8.2
1.0
0.3-19.9
2.0
0.4-365.0
5.0
0.3-59.1
20.5
0.3-16.4
1.8
0.1-134.3
7.5
0.06-6.7
1,3
0.15-9.5
1.2
4.7-1002
34
0.3-57.4
6.8
Indoors
Day
0.4-29.1
4.0
0.03-6.7
0.3
0.4-13.2
1.4
0.3-11.4
1.5
0.5-257.2
5.5
0.5-84.8
14,0
0.3-106.7
3.3
0.1-14.9
2.8
0.06-5.8
0.8
0.2-20.1
0.3
3.4-4230
78.0
0.8-427.1
12.5
Outdrs
Nicfat
0.8-24.8
2.4
0.03-1.0
0.3
0.3-3.5
0.9
0.2-2.3
0.5
0.3-167.3
4.0
0.3-33.1
6.0
0.2-24.0
1.6
0.1-4.7
1.2
0.06-14.9
0.6
0.1-1.6
0.4
3.0-1808
75.0
0.3-114.0
6.4
Outdrs
Day
0.3-190.6
4.7
0.03-0.9
0.3
0.4-6.4
1.2
0.3-2.3
0.5
0.4-66.7
4.0
0.4-18.2
4.3
0.3-57.4
1.8
0.1-27.1
1.3
0.06-4.5
0.5
0.2-1.6
0.4
3.7-1053
68.0
0.3-10.9
2.5
Fixed
Site
0.3-36
3.4
0.2-4.2
1.3
0.0-1.6
0.4
5.6-510
33.1
0.8-37
4.5
a Values shown were not corroborated by GC-MS analyses; an unknown aarpourri
or ocrapounds nay have ooeluted with vinylidene chloride.
146
-------
TAEIE lib
Summary Statistics for Selected Ctsnpounds
by Monitoring Area, location, and Twelve-Hour Period
in ug/m3
PARKVIIIE-OVERIEA
impound
Benzene
Carbon Tetra-
chloride
Qilorabenzene
Moroform
(KJidiLorobenzene
P^ichlorobenzene
Ethylbenzene
Methyl chloroform
Itoachloroethylene
Irichloroethylene
vinylidene
Chloride3
^lene
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Indoors
Nioht
0.1-181.3
21.0
0.01-6.4
1.3
0.2-72.2
4.8
0.1-78.3
9.6
0.1-2893
6.8
0.1-62.8
62.5
0.9-3429
58.0
0.04-508
41.5
0.05-45.6
5.8
0.06-245
5.0
7.7-802.4
107.0
3.7-105.6
27.0
Indoors
Dav
0.9-92.8
14.0
0.04-8.4
1.3
0.2-30.0
3.8
0.9-131.6
7.5
0.4-77.5
4.3
0.2-73.7
6.0
0.3-165.7
9.6
0.1-601.8
39.5
0.02-36.6
4.8
0.1-13.5
1.6
5.0-665.1
86.0
1.4-123.0
18.0
OutdTS
Nidht
0.2-63.4
11.0
0.02-2.2
0.3
0.2-33.4
1.8
0.1-3.7
0.2
0.1-214.9
9.0
0.1-154.5
15.0
0.6-87.4
5.5
1.6-40.9
9.5
0.05-33.1
2.1
0.06-5.0
0.4
5.0-685.4
106.0
1.4-86.5
15.0
Outdrs Fixed
Dav site
0.7-73.76 0.8-19.2
5.0 5.2
0.00-2.3
1.0
0.1-12.1
1.6
0.1-86.5
1.3
0.1-37.3
9.0
0.1-284.3
0.4
0.2-46.1
4.0
0.03-187 1.2-15
11.0 5.6
0.08-12.9
1.7
0.06-5.6 0.2-27.8
0.3 2.1
3.7-428.4
76.0
0.3-70.1 3.0-241
12.0 68.3
1 Values shown were not corroborated by GCMG analyses; an unknown compound
or ocmpounds may have coeluted with vinylidene chloride.
147
-------
TABLE III
CORRELATION BhttWEEN FIXED AND BACKGROUND
Spearman Correlation Coefficients
Compound Dundalk Parkville-Overlg§
Benzene 0.25 0.70
Methyl chloroform 0.20 0.55
Tetrachloroethylene 0.10 0.23
Vinylidene chloride 0.02 0.22
m-Xylene 0.43 0.45
o-Xylenea 0.42 0.46
a Includes styrene.
TABIE IV
CORRELATION HtilWEEN OUTDOOR CONCENTRATIONS OF SELECTED
AT DUNDALK AND PARK7ILLE-OVERLEA
Spearman Correlation Coefficients
Compound Nicdit Day
Benzene 0.4 0.2
Carbon tetrachloride 0.2 0.2
Chlorobenzene 0.15 -0.15
Chloroform -0.25 -0.05
Ethyl benzene 0.45 0.35
Methyl chloroform 0.3 0.2
Tetrachloroethylene 0.35 0.4
Trichloroethylene 0.07 0.3
Vinylidene chloride 0.00 -0.05
m-Xylene 0.3 0.25
148
-------
tl. £ARY R£SULTS OF THE BALTIMORE TEAM STUDY
AIR AND BREATH MEASUREMENTS
A
'
, w. c. Nelsonb, A. Manalec
of Research and Development
gton, DC
P
Off j50riffiental Monitoring Systems Laboratory
fcese e °f Research and Development
c arch Triangle Park, NC
of Program Planning and Evaluation
gton, DC
Baltiroore TEAM Study was undertaken to extend our
of personal exposures, indoor and outdoor air
^ations, and body burden of toxic chemicals. The goals
y design are described in Part I of this paper (Nelson e±
nitor>- S Volume) • Results of the indoor and outdoor air
ririg are described in Part II (Manale et al . . this volume) .
res
u subject carried a Tenax personal monitor for two
a e 12~hour periods (the overnight period, nominally 6pm-
j the daytiltie period, nominally Sam to 6pm) . At the
, middle, and end of the 24-hour monitoring period, each
Provided a breath sample. Subjects filled out one
a °tht aire detailing their household characteristics and
Uv* <24"hour recall) questionnaire concerning their
es on the day they were monitored.
149
-------
Air exchange rates were measured for each 12-hour monitoring
period in each home, providing a daytime and nighttim
measurement. The method employed was the PFT (perfluorotracer]
technique developed at Brookhaven National Laboratory. In thii
method, tubes emitting one of several fluorocarbons at a knovt
rate are placed in a home one or more days before the home is
monitored, to allow for complete mixing. Collector tubes of
charcoal are then placed in the home by technicians during the
first arid second visit and picked up during the second and third
visit. The amount of gas adsorbed on the charcoal tube is
determined by gas chromatography. The average 12-hour ait
exchange rate can then be determined from a knowledge of the
house volume and the time of exposure.
RESULTS
Quality Assurance
Blank levels were consistently low (<10 ng/cartridge) for
nearly all chemicals. Even benzene, normally the worst
contaminant on Tenax cartridges, varied between 9 and 13
ng/cartridge. Recovery efficiencies for all breath samples were
consistently good, ranging between 80 and 115%. Recoveries were
low for one mass spectrometer during one 14-day period, during
which roughly half the personal air samples were analyzed. This
may result in somewhat increased variability for the personal air
measurements. The relative standard deviations (RSDs) of the
duplicate samples also showed increased median values of 30-50*
compared to previous TEAM Study RSDs of 15-30%.
Percent Detected. The number of personal air and breath samples
exceeding the limits of detection (LOD) (about 0.2 ug/m3 for most
chemicals at Dundalk, about 1 pg/m3 at Parkville) is given in
Table I. Since the LODs were so different at each site, the two
sites cannot be directly compared on the basis of the number of
samples detected. That is, the larger number of samples above
the LOD in Dundalk may simply be due to the lower LOD in Dundalk.
Concentrations in Air and Breath
Geometric mean concentrations for samples exceeding the LODs
are shown for air (Table II) and breath (Table III) for Dundalk
and Parkyille-Overlea separately. Concentrations appear to be
similar in the two areas, although it must be remembered that
different limits of detection make comparisons difficult. The
low values for m+p-xylene in air samples in Parkville are
questionable—normally these isomers are 2-3 times the
concentration of ethylbenzene or o-xylene.
Air Exchange Rates
A total of 277 air exchange rates were determined, out of
310 possible (89% completeness). Of these, 12 were so low (0.00-
150
-------
been n? that it was Judged likely that the collectors had not
saae jProperly uncapped. Two additional samples, both from the
conSid°Use' Were at sucn high values (> 1,000 ach) that they were
135 fr^red faulty. This left 263 values, 128 from Dundalk and
Parkville-Overlea. Parkville-Overlea homes had higher
e rates than Dundalk homes (median value of 0.62 h"1
0 °*35 h-1) • This may be due to the larger number of
2lolnes *n the Parkville area. Detached homes would be
to have higher air exchange rates than would row houses.
cllemical at highest levels in both air and breath was
*a natural terpene found in citrus fruits and a popular
a foods, beverages, and lemon-scented products. The
Catcinoer T,°xic°logy Program has recently completed a two-year
Pleariv c. study of d-limonene, with the result that it was
carcinogenic for one of the four sex-species combinations
orrelation between air and breath was noted for
ne *
b» ^ ' ^ndicating that exposure through food and/or beverages
e important.
g tath levels were extremely stable for most individuals
be ^ samP1;i-n9 period, indicating that breath measurements
excellent way to characterize recent exposure
°ns w^^ a^-r exposure were significant for all prevalent
excePt limonene and chloroform, due probably to the
of water and beverage intake for these chemicals.
t their urS had nearly 1° times the benzene levels of nonsmokers
^° 1.5 "teath (geometric mean of 14 jag/m-* for smokers, compared
u naU
o ulta A*' Wallace, L. and Nelson, W. , (1988) "Preliminary
^stero the Baltimore TEAM Study: III. Indoor and Outdoor
Measurements", this volume.
T >
L*A. , Pellizzari, E.D., Hartwell, T.D., Perritt, R. , and
R'' (198?) "Exposures to Benzene and Other Volatile
-27 from Active and Passive Smoking" Arch. Environ. Health
151
-------
CATION OF ENVIRONMENTAL TOBACCO SMOKE
2 PHASE MARKER COMPOUNDS USING
'SIGNAL GAS CHROMATOGRAPHY
L
on
ElM>,r.,,n!eivta3- Monitoring Systems Laboratory
y.S. S*?ntal ^search Center
n Protection Agency
Wangle Park, NC 27711, U.S.A.
]011 ens^onal 8as chromatographic procedure was developed for
of Particulate matter for the tobacco alkaloid, cotinine, and
marker compounds. Analyses were conducted of air samples
controlled laboratory smoking experiments and in private
vit *-fia-rette smoke particulate samples were prepared by extrac-
opa methylene chloride using an ultrasonic probe, concentration by
recons'ti'fcu'bion ir» benzene, and re -con cent rat ion prior to
concentrated extracts were injected on-column into a multi-
^aiy 8as chromatograph. Heart-cuts from a pre-column were
°n a tra-pping column and then transferred on-line to two parallel
ij Columns* Unambiguous identification of cotinine and other
tiffl Vas Stained from the precise, reproducible measurements of reten-
f hg a nit Observed vith two analytical columns of different polarities
?^<1 tr°gen selective detector. Similar cotinine concentrations were
*« "e particulate samples form five different popular brands of
c of the indoor air samples showed that cotinine in a
e etlv^ronment can be readily detected and measured. Results support
a Potential marker compound for environmental tobacco smoke
matter.
155
-------
Introduction
The growing concern over the health risks presented by environmental
tobacco smoke (ETS) is reflected in recent reports of the World Health
Organization1 and the Surgeon General of the United States.2 The 19^
report of the Surgeon General cites involuntary smoking as a cause of lung
cancer and other diseases and, in particular, links it to respiratory
infections in children of smoking parents.
Lewtas and co-workers^ found that the nutagenicity associated with
the concentration of particulate matter in various homes was strongly cor-
related vith the number of cigarettes smoked. Spengler and co-workers'
have found ETS to be an important source of total suspended particulate
matter (TSP) in homes and other indoor environments.
In order to identify and quantitate the contribution of ETS to the
total particulate mass in indoor environments, suitable marker (surrogate)
compounds are required. Nicotine has been used as a marker for ETS (pri-
marily for whole smoke), since it is unique to tobacco and is a major
component of tobacco smoke. However, because of unresolved questions
concerning volatility and reactivity, there are some problems with the use
of nicotine as a marker for ETS particulate matter.5 Consequently, it
would be prudent to look for other potential marker compounds for ETS parti-
culate phase exposure.
Cotinine, another tobacco alkaloid, is less volatile than nicotine
and it, too, can be used with nitrogen specific detectors for enhanced
discrimination from the many other components of particulate matter. A
literature search revealed no sources of cotinine in ambient air other than
tobacco smoke.
Experimental Methods
Instrumentation
Analyses were performed with a Siemens SiChromat-2 multidimensional
gas chromatograph (MDGC) system employing fused silica capillary columns
and equipped with a Varian temperature programmable on-column capillary
injection (Figure 1). Measurements were made with a Perkin-Elmer, LCI-100,
computing integrator. A 1m length of de-activated fused silica tubing
served as a retention gap column and protected the pre-coluran from invola-
tile components in the injected sample. The RSL-200 pre-column was 15ra x
0.32mm ID x 0.5 urn film thickness (Alltech Associates). The DB-5/trap-
ping column was 0.75m x 0.32mm ID x 1.0 um film thickness (J&W Scien-
tific). The DB-5 analytical column was 12jn x 0.32mm ID x 1.0 urn film
thickness (JW Scientific). The RSL-300 analytical column was 15m x 0.32nm
ID x 0.5um film thickness (Alltech Associates). A length of deactivated
fused silica tubing, 0.5m x 0.25mm ID, served as the transfer line between
the colunn switching valve and the flame ionization detector.
Procedures
Sampling. Particulate samples from environmental chambers were
collected with personal sampling pumps operated at 1.7 1pm using untreated
polytetrafluoroethylene (PTFE) filters. Some were also collected with a
denuder sampling assembly (2 1pm) of the type developed by the Harvard
156
-------
School of Public Health. The assembly consisted of a citric acid-coated
denuder tube followed in succession with an untreated PTFE filter and a
citric acid-treated glass fiber filter. The sampler used for residential
air consisted of an untreated quartz fiber filter followed by an XAD-1*
sortent cartridge, and operated at 8 1pm.
Sample Preparation. Particulate samples collected with personal
saapling pumps and those collected on the untreated PTFE filter of the
denuder sampler were extracted with a benzene/methylene chloride (3:1)
aixture using either an ultrasonic probe or an ultrasonic bath. The ex-
tracts were concentrated by evaporation with a stream of charcoal-filtered
helium at room temperature. Particulate samples from the acid-treated
filters and the vapor samples collected in the denuder tube were extracted,
with minor modifications, according to a procedure developed by the Harvard
School of Public Health. The citric acid treated filters were extracted
with a mixture of ethanol, 10 N NaOH, and ammoniated UV grade haxane using a
magnetic stirrer. The vapor phase components were extracted from the
denuder tube with a mixture of sodium bisulfate and ethanol. 10 N NaOH was
added to the extract and nicotine, cotinine, and other basic compounds were
extracted with ammoniated hexane. Nicotine analyses were conducted on
unconcentrated filtered extracts analyses for other compounds were conduct-
ed on concentrated filtered extracts.
The quartz filter and XAD-1* cartridge samples of residential air were
collected and prepared by Battelle, Columbus personnel. Soxhlet extractions
were carried out with methylene chloride for l6h followed by an additional
8h with ethyl acetate. The extracts were combined and concentrated by
Kuderna-Danish evaporation to a volume of 1 ml.
Gas Chromatographic Analysis. The following programs were employed.
Injector: to = 608C, hold for SOsec, ramp to 250°C at 20'C/min; pre-column:
to = 90*C, hold for 3 min, ramp to 250*C at 15'C/min; analytical columns:
to = 50*C, hold until conclusion of cut, ramp to 250"C at 15'C/min. The
helium carrier gas flow rate was 29.6cm/sec at the initial GC temperature
settings, lul ammoniated benzene was co-injected with each sample to
minimize losses of nicotine and cotinine on active sites in the GC system.
Analytes cut from the pre-column were collected on the trapping column and
then transferred to the two analytical columns during the ramping of oven
2, Nicotine and cotinine were identified by their retention times on the
two analytical columns. Confirmation was obtained by GC/matrix-isolation
FTIR analysis. Quantitative measurements were based on peak heights from
DB-5 column chromatogram.
Results
A single GC retention time measurement generally is not sufficient
information to permit unambiguous identification of analytes in complex
mixtures. In this study increased confidence in the identification of
nicotine and cotinine chromatographic peaks was achieved by using a nitro-
gen-specific detector together with two independent retention time measure-
ments. Independent retention times were obtained by the use of two analyti-
cal columns with different polarities. An example of the chromtograms
obtained for nicotine and cotinine in neat extracts of air samples is
presented in Figure 2. The GC procedure employed not only provides com-
plete separation of both pairs of nicotine and cotinine peaks but also
benefits from the economy of using a single detector for both analytical
columns. Precise and reproducible retention times are necesary for identi-
157
-------
fication of analyte peaks. In this study the cotinine peak retention tic
for all cigarette smoke and room air samples agreed with the daily standar:
sample to within 0.008 min.
A fairly consistent ratio between a tobacco smoke constiutent and tti
pollutant of concern (particulate mass) is one of several criteria for i
candidate marker compound reported by the National Research Council. 5 Tbt
results obtained in the smoking chamber studies indicate that cotinine
meets this criterion. In a 1985 study by the John B. Pierce FoundatiK
particulate samples were collected from a chamber containing several activs
smokers using 5 popular brands of filtered cigarettes. The mean cotinine
concentration found in the particulate matter was 7^8 ng/mg, with a relative
standard deviation of 11%. In a 198? chamber study by the John B. Pierce
Foundation a cotinine concentration of 655 ng/nig was found in the particular
matter. Hammond and co-workers6 found that the ratio of nicotine (another
candidate marker compound) to the total ETS particulate matter did not vary
appreciably among k cigarette brands tested.
In the 1987 chamber study the phase distributions of nicotine and
cotinine were explored using denuder samplers. The vapor phase cotinint
collected in the denuder tube was <0.04 ug/m3, which constitutes on]j
10% or less of the total cotinine in the ETS (Table 1). In contrast, vapor
phase nicotine constituted about 78% of the total nicotine in the ETS,
While 84% of the particulate phase cotinine was collected on the untreated
PTFE filter, most of the particulate phase nicotine evaporated from the
untreated filter and was collected on the acid-treated back-up filter. A
comparison of the nicotine and cotinine denuder data indicates that both
alkaloids can claim an advantage serving as a marker compound for particulate
phase ETS. Nicotine offers greater detectability at low smoking rates
"because it occurs at a higher concentration. Cotinine offers a sin^ler
particulate sampling methodology, since prior separation of the (minor)
vapor phase component is not necessary.
The effect of aging on the composition of ETS is another important
consideration in the selection of candidate marker compounds. Nicotine was
stored under air in three Erlenmeyer flasks for periods of up to 68h to
determine if appreciable cotinine might be produced by oxidation of the
nicotine. At the end of the exposure the flask contents were dissolved in
benzene and analyzed. The mean value of the cotinine produced from a
66-68h exposure of 100 ng nicotine was 8.5 pg (±6.3, SD), which repre-
sents only 0.009% of the nicotine in the flasks. Since the average nico-
tine concentration found in the denuder samples was 1*1.9 Pg/m^, the
amount of cotinine which would be produced from the nicotine after 6*8h is
calculated to be only O.OQl* ug/nP. This amount represents only 1% of
the total cotinine found and suggests that the amount of cotinine produced
from air oxidation of nicotine during aging of ETS is negligible.
Another criterion for an ETS marker compound identified by the Nation-
al Research Council is that the compound be present in tobacco smoke in a
quantity sufficient for easy detection, even at low smoking rates. The
cotinine peaks shown in Figure 3 came from a sample collected in a home
where the smoking rate was reported to be 1.3 cigarettes/h during the sampl-
ing period. The injected sample contained 371 pg cotinine. From calibra-
tions conducted between 0 and 0.2 ng cotinine the limit of detection (at 3x
noise) was found to be 15 pg. Therefore, it appears that there is suffici-
ent cotinine in tobacco smoke to serve as a marker for ETS particulate
matter at smoking rates one order of magnitude less than that represented
by the sample illustrated.
158
-------
COIHD results of this study support cotinine as a potential marker
^nin Un(* *"°r env^roninenal tobacco smoke partlculate matter. Cotinine is
It v S t0 totacco smoke and occurs principally in the particulate phase.
in , s found to occur in a fairly consistent ratio to ETS particulate matter
aino Un^e:r studies involving 5 popular brands of cigarettes. No appreciable
°ti °f Co't'inine was produced from nicotine exposed to air for 3 days.
stU(5 ne can be easily detected in indoor air at low smoking rates. Further
are needed to establish the validity of cotinine as a marker compound
variety of environmental and smoking conditions.
B, pj am indebted to Dr. B. P. Leaderer and Dr. P. M. Boone of the John
rf;e F°undation Laboratory and Dr. R. Tosun, formerly of the John B.
*°uridation Laboratory for the cigarette smoke particulate samples
e * J* C* Chuan6 of Battelle, Columbus Division for extracts of
ivirQ al air samples. I am also indebted to R. K. Stevens of the U. S.
nraental Protection Agency for the denuder samples.
ai"ticle has not been subjected to Agency review and does not
reflect the views of the Agency. Mention of trade names or
Products does not constitute endorsement or recommendation for
. «c
Controlling the Smoking Epidemic. Report of the WHO Expert Committee
636 t klnS Control»" World Health Organization Technical Report Series
»*y i 9 ) »
2>
s e Health Consequences of Involuntary Smoking. A Report of the
G pgeon General," U.S. Department of Health and Human Services, U.S.
°vernment Printing Office, Washington, DC (1986).
T
» Levtas, S. Goto, K. Williams, J. C. Chuang, B. A. Peterson, N. K,
Uson, "The Mutagenicity of Indoor Air Particles in a Residential
*lot Field Study: Application and Evaluation of New Methodologies,"
SSaSli-Bnviron. 21:1*1*3 (1987).
' j ~~
p* D» Spengler, D. W. Dockery, W. A. Turner, J. M. Wolfson, B. G.
pj7is»Jr. , "Long-Term Measurements of Respiratory Sulfates and
articles Inside and Outside Homes," Atmoa. Environ. ,15:23 (1981 ).
B
°nal Reaearcn Council, Environmental Tobacco Smoke* Measuring
and Assessing Health Effects, National Academy Press,
, DC (1986).
" K* Hammond, B. P. Leaderer, A. C. Roche, M. Schenker, "Collection
alysis of Nicotine as a Marker for Environmental Tobacco
" Atmos. Environ. 21:1*57 (1987).
159
-------
Table I. John B. Pierce Foundation environmental chamber
denuder sampler
CONCENTRATION IN AIR (yg/m3)
COTININE
DENUDER TUBE
TEFLON FILTER
TREATED QUARTZ FILTER
NICOTINE
DENUDER TUBE
TEFLON FILTER
TREATED QUARTZ FILTER
DENUDER #1
12/3/6?
O.Ol*
0.31
0.06
0.9
9.7
DENUDER #2
NOT DETECTED
0.38
NOT ANALYZED
31
2.1
6.0
160
-------
On
to
— Column
ijector
FID
Retention
Gap
Column
Union
\
Splitter r
Union L
c
50000
F
i
RSL-200 Column
Pre-Column Switching
Valve
NSD
^ommmJ
DB-5 Analytical Column
RSL-300 Analytical Column
JDB-5
Trapping
Column
J
OVEN I
OVEN 2
Figure 1. Schematic of the flow paths 1n the multidimensional gas chromatograph.
-------
1
19
20
21
22 23
TIME (Min)
24
25
26
Figure 2. NSD chromatogram of the combined nicotine and cotim'ne
heart-cuts from the concentrated extract of an ETS
particulate sample.
162
-------
in
I
m
o
I
I
I
20 21 22 23 24
TIME (Min)
25
26
3. USD chromatogram of the cotlnine heart-cut from the
concentrated extract of a living room air sample.
163
-------
DESIGN OF A GLASS IMPACTOR FOR AN
ANNULAR DENUDER/FILTER PACK. SYSTEM
P, Koutrakis, J.M. Wolfson, M. Brauer, and J.D. Spengler
Harvard School of Public Health
665 Huntington Avenue, Boston, MA 02J15
R.K.. Stevens
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
A glass impactor for an annular denuder/filter pack system was developed, to fu
the application of denuder technology in sampling atmospheric gases and particles.
glass impactor consists of an entrance section containing the inlet tube, the accele
jet, and the impaction plate which is mounted at the entrance to the annular denuder
impaction plate is a removable porous glass disk which can be impregnated with miner
to avoid bounce-off of the collected particles during sampling. Calibration tests sho .
that the impactor has a 50% aerodynamic particle cut-off point of 2.1 /im, at a
10 Lmin'1. Particle loss experiments were conducted. Losses on surfaces insid* J.
impactor, annular denuder, and filter pack, determined for particle sizes ranging l>et*'tf
1.50 and 2.77 /xm, were found to be very low in each of the sampler sections, with
losses lower than 3%.
164
-------
PRODUCTION
Th
analys- e measurement of trace atmospheric gases relies on a variety of sampling and
sUrfa techniques. Of particular interest is the collection of trace gases on denuder
these ! Which are specially coated for selective adsorption of different gases.1-2 However,
3 LJJJ. _"dies used cylindrical diffusion denuders, which require flow rates less than
C]uant-. .' Higher flow rates, 10-20 Lmin'1, require considerably larger tubes to obtain
We collection efficiencies. Other studies used a parallel connection of denuder
w use of increased flow.3-4 Denuder technology was advanced through the
o .
air Da .ftnt °^ a more versatile denuder design, consisting of two coaxial cylinders with
effjc- !ng through the annular space.5 The annular denuder achieves collection
cyc]0 ies c'°se to 100% at laminar flows up to 20 Lmin'1. Investigators have employed
the an °f otner metallic impactors to remove coarse particles, which can be deposited on
the us« ^ denuder wa"s and thus cause overestimation of gas concentrations. However,
enter t, *nese particle pre-selectors can result in a significant loss of gases before they
The QK- , nuder series. Such losses are minimized using a newly designed glass impactor.
^arao*. ^tlve °^ 'his Paper is to describe the glass impactor and present the results of its
cterization.
II>T1ON OF THE GLASS IMPACTOR
Th
one gj Sam£ling system, shown in Figure (1) consists of a borosilicate glass impactor,
an entr anr»ular denuder, and a virgin Teflon (FEP) filter pack. The impactor consists of
( 0^ e'utr'ator section containing the inlet tube followed by an acceleration jet, and
plate- The plate is mounted at the entrance to the first annular denuder.
1.3 Cm riat<>r section is 9.5 cm in length, with i.d. of 1.1 cm. The acceleration jet is
w'th a °n8' w'th an i-d. of 0.300 + 0.005 cm. The impaction plate is a porous glass disk,
fate isn°minal Pore size of 10-J6 ^m, diameter of 1.1 cm, and thickness of 0.16 cm. The
fficti0n ^Ul)ted in a removable virgin (FEP) Teflon holder, which is securely attached by
deii(ier U into a cyJJndricat glass cavity fused to the entrance of the first annular
-------
1MPACTOR CALIBRATION AND PARTICLE LOSS TESTS
Impactor calibration tests were conducted at the University of Minnesota, using the
procedures of Marple and Rubow.7 Two samplers, designated as impactor #1 and impactoi
#2, each consisting of a glass impactor, an annular denuder, and a filter pack, were tested.
Experiments were performed on both impactors to determine aerodynamic particle collection
efficiencies and particle loss characteristics. A vibrating orifice monodisperse aerosol
generator was used to produce uranine-tagged oleic acid liquid particles. Collection
efficiencies and particle losses were determined by extracting the test particles from the
impaction plate, filter, and interior surfaces of each sampler with an aqueous solution,
measuring the fluorescence of the extracts. Flow rate was maintained at 10 Lmirr1, using
Millipore critical orifices.
Figures (2) and (3), showing the aerodynamic particle collection efficiency curves and
calibration data for impactors #1 and #2, reveal that both impactors have a 50%
aerodynamic particle cut-off point of 2.1 /;m, have very sharp cut-off characteristics, and
are in close agreement. Table (1) summarizes the results from the particle loss tests.
Surface Josses for particle sizes ranging from 1.50-2.77 /im, were measured inside the inlei
tube, around the outside of acceleration nozzle, inside the annular denuder, and inside the
filter holder. These losses were found to be very low in each of the sampler sections,
with total losses lower than 3%.
MEASUREMENT OF FINE PARTICLE MASS AND SULFATES
Pilot air sampling experiments were conducted on the roof of the Harvard School of
Public Health in downtown Boston, MA during the summer of 1987. Three samplers,
consisting of the glass impactor, two annular denuders, a filter pack, and a flow-controlled
pump operating at 10 Lmin'1' designated as the Harvard/EPA Annular Denuder System
(HEADS), were co-located three Harvard impactors (HI). The HI system has been designed
and characterized to have a 50% aerodynamic particle cut-off of 2.5 pm and a flow rate of
4 Lmin-1.8 Sample duration varied between one and three days depending on the observed
air quality levels. Mass concentrations obtained from the HEADS samplers were
consistently about 10% lower than mass concentrations determined by the co-located HI
samplers. This can be explained by the slightly lower aerodynamic cut-off of the HEADS.
Next, a comparison of the concentrations of sulfate collected on the teflon filter for
HEADS and HI systems showed excellent agreement.
CONCLUSIONS
A glass impactor was developed for an annular denuder/filter pack system.
Calibration tests showed that the glass impactor has a 50% aerodynamic particle cut-off
point of 2.1 pm, at a flow rate of 10 Lmin'1. Particle loss tests were conducted for two
samplers. Losses on impactor, denuder, and filter pack surfaces, determined for particle
sizes ranging from 1.50-2,77 /im, were found to be very low in each of the sampler section
with total losses lower than 3%. Aerosol fine particle mass concentrations determined
using the glass impactor were about 10% lower than those from the Harvard impactor, while
sulfate concentrations were in excellent agreement. The difference in mass concentrations
is due to the slightly lower aerodynamic particle cut-off of the glass impactor.
166
-------
ACKNOWLEDGEMENTS
W°rl< was supported by the National Institute of Environmental Health Sciences
C0ntac 8rant NOS. 1R01 ES0495-01 and ES-07 155-03, Electric Power Research Institute
Envir RP-1001, and the Department of National Health and Welfare Canada,
of (|j T ^ntel Protection Branch. We would like to acknowledge Professor Virgil A. Marple
to E)r » lversity of Minnesota for the characterization of the impactor. A special thanks
wis t Im Slater and Gerald S. Keeler for their invaluable contributions. Finally, we
XPress our appreciation and thanks to Larry Stone, University Research Glassware
for fabricating the glass impactor.
^NCES
Co .am> J-L., Wilson, W.E., and Bailey, E.B., Application of an SO2-denuder for
]•)< o, °Us measurement of sulfur in submicrometric aerosols," Atmospheric Environment.
2" *33 (1978)
13. ,''•> "Method for determination of atmospheric ammonia."Atmospheric Environment.
1 IJ85 (1979).
i
' ^teven t> v TX
atmo ' . Dzubay, T.G,, Russworm, G., and Rickel, D., "Sampling and analysis of
Pheric sulfates and related species," Atmospheric Environment. 12: 55 (1978).
' F°rest j c
njtr t * ' ^Pandau, D.J., Tanner, R.L., and Newman, L.,"Determination of atmospheric
Env' a nitric acid employing a diffusion denuder with a filter pack," Atmospheric
"^EQflIQSni, 16: 1473 (1982).
5. j>Qs
t]^e in'» M.» Febo, A., and Liberti, A., "New design of a high-performance denuder for
^PUng of atmospheric pollutants," Ajtm^jpj^£igJEjrwij[ojLme_ntt 17: 2605 (1983).
» Y'A. and Rubow, K.L., Cascade Impactor Sampling and Data analysis. American
Assoc. Monograph series, pp 79-109.
V.A, and Rubow, K.L. Development of a microorifice uniform deposit cascade
Final report for the U.S. Department of Energy, Pittsburgh Energy
V Center. Contract DE-FG-22-83PC61255 (1984).
-"Pa ^'^'' ^ubow« K-L-» Turner, W., and Spongier, J.D., "Low flow rate sharp cut
3?- ,,„ rs *°r indoor air sampling: Design and calibration," Jour. Air Poll. Contr. Assoc..
' 13°3(1987).
167
-------
Table 1: Particle Loss Results
Particle Size
(um)
Particle Loss, Per Section
Impactor #1
234
Impactor #2
1 2 3
1.50
1.75
2.00
2.19
2.40
2.77
0.3
0.4
0.1
0.1
0.1
0.8
0.3
0.4
0.1
0.1
0.5
0.1
0.7
1.1
0.7
1.0
0.5
0.8
1.0
1.1
0.2
0.1
0.1
0.8
0.5
0.3
0.5
0.1
0.6
0.5
0.3
0.2
0.1
0.6
0.9
0.6
0.9
0.7
0.6
_^
Partide losses on
surfaces inside inlet tube
Particle losses on
surfaces around the
outside o( nozzle
^Inlet
IMPACTION
PLATE
Panicle losses on
inside surfaces ol tube
\
Denuder tube
Particle losses on
inside surfaces of
Mlor holder
u
„ FtTER
\Filter holder
Figure I:
Personal Sampler
168
-------
COMPARISON OF METHODS FOR MONITORIMG
DRY DEPOSITION POLLUTANTS: SUMMER 1987 STUDY
E. Hunter Daughtrey, Jr., David K. Bubacz, and Dennis D. Williams
Northrop Services, Inc. - Environmental Sciences
Research Triangle Park, North Carolina 27709
William A. McClenny and James D. Mulik
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
A study was conducted at the Dry Deposition prototype station at the
EPA Annex, September 2 through October 20, 1987, to compare the candidate
methods for monitoring acidic deposition species. The methods under
study were the annular denuder system (ADS), the transition flow reactor
(TFR) concentration monitor, and the Canadian Filter Pack (FP). Weakly
integrated samples were taken for seven weeks by the three methods, and
daily samples were collected for two of the weeks. All analyses were
performed by ion chromatography.
Comparisons were made of operational factors for each method and of
the quantitative results obtained. Most of the operational
considerations centered on the ADS, because the components used were of a
recent design. Design improvements included the impactor inlet and a
better machined filter pack. Deficiencies due to rigid connectors
included proneness to leaks and breakage. Comparison of quantitative
results was made among methods and between daily and the corresponding
weekly results. Interconversion of nitrate-related species, such as
oxidation of MONO to HM03 and the volatilization reaction
NHi|N03(s)^NH3(g) -*- HN03(g), confounds the comparison between daily and
weekly results. Values for total nitrate mass balance showed reasonable
agreement between daily and weekly results and, to a lesser degree, among
methods. Results for other species either showed reasonable agreement or
were consistent with results obtained in the 1986 study.
170
-------
1987 STUDY °F METHODS FOR MONITORING DRY DEPOSITION POLLUTANTS: SUMMER
°ctober St?n y was conducted at the EPA Annex parking lot, September 2 -
c°noentr f ^987, to compare methods used to measure ambient air
The neth °^ species of interest in the dry deposition processes.
transiti under study were the annular denuder system (ADS), the
fiiter D°n. flow reactor (TFR) concentration monitor, and the Canadian
^Uantitat? ^FP^' Comparisons were made of both operational features and
fUl in ,, results. The purpose of this study was threefold: (1) to
pP), (g Saps" from the 1986 study! (i.e., weekly data for the TFR and
to test the methods and protocols with new operators, and (3) to
complete understanding of the processes occurring while
TK
e methods were each run according to the operational
'jsed in the 1986 study to help ensure comparability with that
^ P^cate samples were taken for all weekly and daily
°nty a "ts» with the exception of the daily ADS measurements, for which
p°inatoff °^e syst;era was available. All samples were analyzed by ion
?*CePt f aPky. Ion chromatography was satisfactory for all measurements
6ffei NH3 captured on Nafion® in the TFR because of a coeluting,
nS species. Seven sets of weekly data were taken, as were 14
aal*y samples (in weeks 3 and 5 of the study).
Discussion
^1118 the three methods, the operational features and the
result3 were compared. Operational considerations include
tj e qu ' d*sadvantages, and observations for each method. Comparison
c » asan ative results was made between methods for like periods of
la^sponrn We^ as between weekly integrated measurements and the
to id averaSesi of daily taeasurements. In addition, attempts were
er»tify physical and chemical causes for observed differences.
Operational Considerations
PlP>XjvS£netit ADS* both tne new fiit61" Paok and impactor inlets were
v ^tj,, cnts operationally over the previous design. The new rigid
leatrneCt°rs were viewed as a disadvantage because of a proclivity
anc* breakage of the glass denuder sections with their use.
the cyclone inlet was less than satisfactory.
- ,
h a % nan8eover for the TFR was the most labor-intensive and time-
ttt ^intli0 any reclulred for the three methods. This created difficulty
^Q other 8 comparable sampling times between the TFR method and the
red u111^11^^. This was largely due to the cleaning and drying time
ed u
ar ft? e Protoc°l' Problems in the extraction of the cyclone were
ot* those encountered with the ADS.
he f * 1 1
it pack ^Fp^ waa the ea3lest of tne tnree methods to use.
Q did not provide any particle size fractionation. The
of 8ec' ^n *-ne study did not adequately specify the correct "up"
Zefluor* filter. As a consequence, particulate
171
-------
concentrations determined by the FP are inaccurate, because ultrasonic
extraction is incomplete if sampling occurs on the wrong side of the
filter.
Quantitative Results
Summaries of the quantitative results of the study are grouped into
chemically similar classes: sulfate, which includes sulfate and sulfur
dioxide; ammonium which includes ammonium ion and ammonia gas; and
nitrate, which includes nitric and nitrous acids and particulate nitrate.
Tables 1 and 2 summarize the sulfate results. Particulate sulfate
values agree reasonably between the ADS and TFR methods and between daily
and weekly averages. The FP results are unreliable because of the
Zefluor® filter loading problem. Sulfur dioxide results are lower for
the TFR and FP than for the ADS. Similar observations were noted in the
1986 study and by other workers. 1,2
Tables 3 and 4 give the results for the ammonium group. Particulate
results for ammonium show reasonable agreement among all methods and
between sampling periods. Some loss of ammonia is seen from the ADS
citrate denuder in the weekly samples.
Tables 5 and 6 give the results for the nitrate group. With the
ADS, nitrous acid appears to be slowly oxidized to nitric acid over the
weekly sampling period. Volatilization of ammonium nitrate, indicated by
the volatile nitrate, is also greater for weekly samples. The nitric
acid and fine-particle nitrate results calculated by the TFR protocol
method yield confounding results, because the measured
strip/(strip+filter) ratios are grossly different from the protocol
ratio. Closer agreement to the other methods is obtained if
concentrations are calculated by treating the TFR as a filter pack.
Nitric acid results for the ADS and FP are similar to one another but not
to the TFR results. Particulate nitrate results for the FP are again
confounded.
A brief experiment was performed with the ADS to compare the new
impactor inlet to the cyclone inlet. Results are shown in Table 7.
Differences are largely attributable to the more complete extraction
possible with the impactor. Further study is needed.
Conclusions and Recommendations
• Species interconversions are particularly important for the
nitrate group (e.g., nitrous to nitric acid;
NHi|N03(s);?NH3(g)+ HN03(g)) for all methods.
• TFR results for nitric acid and fine-particle nitrate are
inaccurate. The error can be traced to large differences
in measured and theoretical ratios of nitric acid collected
on the nylon strip to total nitric acid collected.
• Filter pack results are confounded by the Zefluor® filter
problem.
• Preliminary results indicate that the ADS impactor is
superior to the cyclone inlet.
172
-------
The fragility of glass denuder sections with the hard
Plastic connectors is a major operational concern.
Sampling periods should be limited to 24 h or less to
interconversion problems.
conclusions and recommendations given here are those of the
ilnPHeci autnor (END). No endorsement by EPA personnel is intended or
f Sampling and Analytical Methods Development for Dry
Monitoring Report RTI/2823/00-15F, Research Triangle
> Research Triangle Park, N.C., 1987.
D> Kt Butacz> D- D- Williams, J. D. Mulik and
Dry DeP°sition Methods Comparison Study, Report NSI
-0] Northrop Services, Research Triangle Park, North Carolina, ]988.
173
-------
Table 1. Comparison of Weekly
Means "Sulfate" Group
Species
SOaasSC^
PartiailateSOf
C
F
T
Total Sulf ate
ADS
9.94
0.20
8.27
8.47
15.95
TFR
5.59
0.20
8.44
8.64
1423
FP
5.82
6.61
12.43
Table 2. Daily/Weekly Comparison
"Sulfate" Species
Species
SO2asSO;
ParticulateSOf
C
F
T
Total Sulfate
Daily/
Weekly
D
W
D
W
D
W
D
W
D
i W
ADS
7.40
9.50
0.75
0.16
6.48
7.99
7.15
8.15
14^5
1531
TFR
5.55
4.86
022
0.18
739
9.10
7.61
928
13.15
14.14
FP
4.70
6.33
3.88
5.90
8.58
1224
Table 3. Comparison of Weekly
Means "Ammonium" Group
Species
NHjasNH4+
ParticulateNH*
C
F
T
Total Ammonium
ADS
0.29
0.03
1.28
131
1.60
TFR
-
0.03
135
138
1.38
FP
1.53
1.53
Table 4. Comparison of Daily vs. Corresponding
Weekly Means for "Ammonium" Group
Species
NH3asNH+
ParticulateNH*
C
F
T
Total Ammonium
Daily/
Weekly
D
W
D
W
D
W
D
W
D
W
ADS
0.82
0.21
0.03
0.03
122
1.10
1.25
1.13
1.87
134
TFR
—
0.08
0.02
1.54
1.76
1.63
1.79
1.63
1.79
FP
•«
1.36
1.59
136
159
-------
TableS. Comparison of Weekly Means, "Nitrate" Group
Species
F Comments i ADS
TFS
FP
HNQ as
2.07
2.51
LJ7
£45
HOIVO as
0.17
(0.912)
PaiticuUte NOf
C
F
V
T
Ibtal Nitrate
**
4*
»»»
OJ9
O.ftl
0.41
0.82
3.10
0.35
-1.11
0.08
_
-0.76
0.43
1.75
1.80
0.45
3.81
j calculated as NO" (strip+nylon filtw>A'l,
"Fine tiO~ cakuiated a$ NO~ Teflon with no correction for HNO}.
•"Above alternative calculation methods used.
""Estimated from NO" on carbonate filter
Table 7. Comparison of ADS Impactorand Cyclone
Inlets on Found Particulate Concentration
(means of three weeks)
Coarse
Fraction
Impactor/
Cyclone Ratio
Fine
Fraction
hnpactor/
CycloneRatio
Impactor
Cyclone
(I)
(0
(D
(O
NO-
0.688
0.493
1.40
ao3i
0.028
1.11
so;
0340
0.219
1.55
8.130
7.420
1.10
NH+
0.087
0.031
2.81
1327
1^1
1.06
Table 6. Comparison of Dally and Corresponding
Week/y Means, "Nitrate" Group
Species
HNO3 as NO~
HONO as NO"
Particulate NO"
C
F
V
T
- Total Nitrate
NO,
Daily/
Weekly
D
W
D
W
D
W
D
W
D
W
D
W
0
W
D
W
D
W
D
W
D
W
D
W
Comments
»
••
• 4
»•»
ADS 1 TFR f FP
1.00
1.78
1.11
0.19
0,66
OJ8
0.12
0.02
0.1*
0,35
0.91
0.74
3.02
2.71
2.30
2.21
1.76
1.27
-
0.34
0.44
-0.45
-0.92
0.13
0.06
_
-
-0.11
-0.48
0.57
0.50
2.19
1.73
2.33
1.79
2232
18.84
1,27
2-02
(0-77)
0.44
0.44
1.94
3.23
j calculated *s NOS {strip+nylon
**Fine NO" calculated as NO" Teflon with no correction for HNOr
"'Above alternative calculation methods used.
""Estimated tram NO~ on carbonate filtet
-------
MEASUREMENT OF ATMOSPHERIC AEROSOL ACIDITY:
LOSSES FROM INTERACTIONS OF COLLECTED PARTICLES
J. L. Slater, P. Koutrakis, G. J. Keeler,
J. M. Wolf son, and M. Brauer
Harvard School of Public Health
665 Huntington Avenue, Boston, MA 02135
While existing methods of atmospheric aerosol acidity measurement adequately prevent
neutralization of fine particle acidity by removing alkaline coarse particles and gaseous
ammonia from the air sample, these techniques do not consider particle interactions on the
collection medium, which may cause underestimation of the aerosol acidity. A quantitative
assessment of acid neutralization due to such interactions is made possible using a newly
designed system. This system includes a glass impactor to remove coarse particles, a series
of denuders to collect acidic and basic gases, and a filter pack. The filter pack contains a
Teflon filter for fine particle collection and a series of three treated cellulose filters to
trap gaseous products from the collected fine particles. The first cellulose filter is coated
with sodium carbonate to trap nitric acid originating from the dissociation of NH4N03
collected on the teflon filter and from the displacement of the sulfate related hydrogen ion
by ammonium nitrate. A second sodium carbonate-coated filter is used to correct for
artifact nitrate on the first cellulose filter. Last is a citric acid-coated filter used to
collect ammonia from the dissociation of the ammonium nitrate on the Teflon filter.
Determination of the differences between moles of corrected nitric acid and moles of
ammonia, allows a quantitative correction for the neutralized acidity on the teflon filter.
Preliminary results from the Harvard Acid Study suggest that large fractions of the aerosol
acidity can be neutralized during collection on filter media.
176
-------
assoc.ring the last two decades, many investigators have measured atmospheric acidity
rain ^^ With rain' Particles, and gases. Although there is a large data base for acid
' n'tr'C acid' few measurements of acidic aerosols have been reported.1 Currently,
measurement techniques require removal of alkaline coarse particles as well as
ammonia Prior to collection of acidic fine particles, to avoid neutralization.2'3-4
y' however' does not consider potential particle interactions on the
media resulting in underestimation of atmospheric aerosol acidity. This paper
of ac-, s a new sample collection scheme, which makes possible a quantitative assessment
neutralization due to such interaction.
DESIGN
^ strong acidity was measured with the Harvard-EPA Annular Denuder System
filte *' The sampling system consisted of a glass impactor, three annular denuders, and
as shown in Figure 1. The glass impactor has been found to have a 50%
Particle cut-off at 2.1 /urn and has been shown to effectively remove coacse
1 a flow of 10 Lmin'1.6 The impactor removes only coarse particles, allowing
fine particles to pass into the annular denuder and filter pack components. The
'ecies HNO3, HNO2, and SO2 are trapped by a Na2CO3-coated annular denuder
of nitrvu °y a second Na2COs-coated annular denuder used to determine artifact formation
the ft 6 and nitrite, for correction of the apparent concentration of HNO3 and HNO2 on
denuder. The third denuder was coated with citric acid to collect gaseous
Laboratory tests of the denuder collection efficiencies and capacities have been
Previously.^
The s
"he fi^16.8 °f f°ur denuders was followed by a filter pack which contained four filters.
cieilc * filt« was a 47 mm diameter, 2 pm pore PTFE Teflon membrane (Gelman
itrate "^ to collect the fine particles for aerosol strong acidity, ammonium, sulfate,
fillip' nd nitrite determinations. The second filter was a 47 mm diameter cellulose filter
Solutiotj trfated with a 2% (w/v) Na2COs and 2% (v/v) glycerol in 3:10 methanol/water
This fiiter is used to trap HNO3 originating from the dissociation of NH4NO3
*n the Teflon filter (equation 1) and from the displacement of the sulfate related
10n by ammonium nitrate (equation 2).
'* "* NH3(g) + HN03(g) (1)
'HS04, (NH4)3H(S04)2 + NH4N03 --> (NH4)2S04 + HNO3(g) (2)
Jle ID.the H£ADS sampler nitrogen oxides and PAN are not removed from the air
nitrart'fact njtrate and nitrite can be formed in situ due to the interaction between
sctio8en °tides and/or PAN with the Na2CO3-treated cellulose filter. To allow
"«vtistr Or this artifact nitrate, a second Na2COs-treated cellulose filter is placed
°ccurrjnam of the first Na2CO3-treated filter. We assume that the artifact reactions
Ol» the f-°n the second Na2CO3-treated cellulose filter are equivalent to those occurring
" ,/fSt Na2C°3-treated filter. Therefore, the nitrate concentration on the first
Lreated filter can be corrected by subtracting the nitrate concentration on the
j a^-°3-treated filter. Last is a citric acid-coated cellulose filter to trap ammonia
dlssociation of NH4NO3 collected on the Teflon filter (equation 1). Using this
177
-------
filter pack system, the apparent aerosol strong acidity measurements on the Teflon filter
can be corrected by adding the moles of corrected nitrate and nitrite from the first
cellulose filter and subtracting the moles of NH3 measured on the citric acid-treated filter.
H* (TOTAL) = H+(FI) + H+(correction) <3)
^(correction) =
-------
aPpare er,va'ue although few winter measurements are available for comparison.4 The
'nteracf ^^'^ on January 16 and 22 was low. The strong acidity losses from particle
the nS ° l^e Teflon filter were again calculated using equation 4. On January 12
Str .
stronc ? acidity correction was 34.6 nmole/m3 which represents 66% of the apparent
Dart of i! ^ Measured on the teflon filter. This one measurement suggests that a large
Previou 6 St!°n^ acidity can be lost by particle interactions and that measurements made
acictity ^ W't*l0ut this technique, may have significantly underestimated aerosol strong
c°rrecf n 'anuarv 16 the apparent strong acidity was negligible while the strong acidity
Particie • WaS significant- 1° this experiment, all of the strong acidity was lost by
stro n act'ons- On January 22 the apparent strong acidity was again very low and
ac'dhy correction was also low. The nitrate and ammonium concentrations were
in showing that the dissociation of NH4NO3, was the important process for that
Period.
occ m'nary data presented in Table 1 suggest that aerosol strong acidity losses
H °n *^e ^e^on filter due to interactions between collected particles. The three
me nuders followed by a four stage filter pack makes possible a quantitative
Dt of acjd neutralization on the Teflon filter.
~"'lt On K^ Was suPP°rted by the National Institute of Environmental Health Sciences
C0ntract s 1R01 ES0495-01 and ES-07155-03, the Electric Power Research Institute
El*virnn nuraber RP-1001, and the Department of National Health and Welfare Canada,
°nn*ntal Protection Branch.
NCES
orv PreciPitat'On Assessment Program (NAPAP) Annual Report, NAPAP,
2. p. * *^C. (1986).
m^P- firachaczek, W.W., Truex, T.J., Butler, J.W., and Korniski, T.J. "Ambient
ttorth«*f* rements on Allegheny Mountain and the question of atmospheric sulfate in
ineastern United States." Ann. N. Y. Acad. Sci. 338:145-173 (1980).
—, T.G., Shaw, R.W., McClenny, W.A., Lewis, C.W., and Wilson, W.E.
/*i!t °f tne aerosol in the Great Smoky Mountains." Envir. Sci. Technol.
j. - •'» (1980).
te,rOsol stm«-> Wolfson, J.M., and Spengler, J.D. "An improved method of measuring
^•'ngston 2,"8 acidity: Results from a nine-month study in St. Louis, Missouri and
^ un» Tennessee.1* Atmospheric Environment 22:157-162 (1988).
a?ne»aCSf P" J^oMson, J.M.. Slater, J.L., Brauer, M.. Spengler, J.D., Stevens, R.K., and
r°sok a'nH Evaluation of an annular denuder/filter pack system to collect acidic
ana gases." Submitted to Environmental Science and Technology (1988).
atW *£0ptrakis, P Slater, J.L.. Wolfson, J.M., Spengler, J.D., and Stevens, R.K.
Jn«t A the Harvard/EPA Annular Denuder System." Submitted to the Proceedings
.am A , Meeting of APCA (1988).
179
-------
IHTRODUCTIOH OF A NO/HO MONITOR FOR THE
SUB-PPB RANGE
Werner Martin
Tecan US, Inc.
P.O. Box 21*85
Chapel Hill, North
Carolina 27515
A cooperative effort between a research institute,
Institute for Chemistry of Perturbed Atmosphere of the
Nuclear Research Center Juelich, Federal Republic of Germany,
and an analytical instrument manufacturer, Tecan AG,
Switzerland lad to a new product. This product is a
chemiluminescence NO analyzer for low concentration
measurements.
The new analyzer CLD TOO ppt has a lower detection limit
of 50 ppt. This alone might not impress too many, but the
fact that all chemical interferences are eliminated by
determining the chemical zero point for each measurement
point, guarantees accuracy otherwise not achievable.
A totally automatic NO/NOx monitoring system is described.
It will produce reliable, highly accurate data down to
concentrations of 50 ppt of NO.
Keywords: Oxides of nitrogen, ppt, interferences,
instrument, monitor, chemical zero point,
automatic, PC-controlled,
chemi luminescenee , controlled,
photolytic converter
182
-------
this 6 devel°pment of the product introduced and discussed in
ketwe ^er Was P088*-10!6 <^ue to a cooperative agreement
of peetl Tecan AG, Switzerland and the Institute for Chemistry
^Ueli ?Urbec^ Atmosphere of the Nuclear Research Institute in
v°rki Federal Republic of Germany. Andreas Volz has been
« a« with us to make it possible to present at the APCA
1988 in Dallas , the first commercially available ppt
/N0x/N02 chemiluminescetice monitoring system.
, he cannot attend this conference. He is also
as a co-author, as noted in the program.
s » Tecan would like to acknowledge his valuable
ution to previous presentations he made with us, and
^ ti K L*
*• nitn for being so dedicated to making this technology
er from his institute to Tecan successful.
to 1982 Tecan's expertise had been in R&D and
J*ob °f a11 sort8 of Photometric, electrochemical
e TO ^nstrumentat ion for analytical laboratories.
Tecan has produced and marketed chemiluminescence
worldwide.
°een a ^f the principles applied in this new instrument have
Heed ln exPeriraental instruments for quite some tlm
Pat>tlcui for the correct measurement of NO and N02 has been
p°llut . arly Pressing in situations where levels of those
e*Uu i are below concentrations of 20 ppb 10-20 ppb.
r tn ence has been the state of the art technology
NO and N02 for more then a decade. In readily
e j cllemiluininescence analyzers no adjustments are made
fac£struinent 's signal for chemical interferences. The
a ' that the sienal SetB adjusted for is the dark
darlc current allows us to subtract noise
- the optical and electrical system. As long as
gp Ses contain NO concentrations above 20 ppb, chemical
g ei}ces are not seriously hampering the accuracy of the
measurements of lower concentrations should not
2~^ PPb interferences, then a chemical zero point
Determined by the instrument.
UQt a4- reaction which produces the fluorescence signal
8 i
-------
that can be made in order to be able to determine the zero
point^1'2!. This phenomenon lead to the idea of determaining a
chemical zero point in a new instrument that has a
pre-chamber and a main reaction chamber.
The Analyzer
The NO-Analyzer OLD TOOppt consists of an ozone genera*0''
a NO-reaction chamber, a zeroing chamber, a photomultiplY6*
tube, an ozone scrubber, valves, pressure and flow
controllers, and a microprocessor. The instrument has two
basic measuring cycles (Figure 3):
Cycle 1: Sample .passes through tubing to the point
the ozone coming from the ozone generator gets mixed in* *
four way valve diverts the sample/ozone mixture to the ,
reaction chamber. A NO first signal is produced. It inclU"
the NO, the dark current and the signal from the chemical
interferences .
Cycle 2: Sample follows same pass as in cycle 1, excep*
the sample/ozone mixture is diverted first into a pre-cfc
and then to the reaction chamber. This allows all NO to
react with ozone before the mix reaches the reaction cha
where other reactions that produce a signal are taking P
By subtracting this signal from the signal produced in Cyc *
1 adjustments can be made for all chemical interferences *
the dark current.
The Converter
In order to measure N02 at low levels an external N02
NO converter is necessary. A photolytic converter has bee?*Ji
developed that consists of a reaction chamber with a UV * ^
source. All significant interferences common to commerci*
available converters are eliminated with this specific
design.
Two additional cycles are added to the NO measuring
determine, in the same fashion, a correct NOx measurement
(Figure k).
The photolytic converter is specifically converting ^ f
and no other components to NO. The conversion efficiency „
the Tecan converter is dependent on the ozone concentrat*
in the sampled air. For continuous monitoring of low .,
concentrations of N02, it is necessary to monitor ozone
the same site in order to correct for the variation of
conversion efficiency.
184
-------
TECAN's CRA-HOX System
^Can's CRA NOX System (Figure 5) represents a complete
evel NO/K02 monitoring system that is totally automated.
elements are the Tecan CLD 700 ppt NO analyzer, the
nai photolytic converter, an ozone analyzer, a gas
ator» a aero air generator, a PC-AT and software. The two
and the titrator have RS 232 communication ports
the software to do data acquisition and controlling.
Applications
The
of j,^, c°raniercial success of such a system is sought in areas
ound level network stations, in research of dry
-°n in laboratories and the field and in atmospheric
;ai-.S from air crafts. There is certainly a need for
y automatic measurement equipment for remote stations
4 8epvice and maintenance cost savings would make a
-ra auch aa Tecan1s CRA NOX a bargain.
Therp
Of HO/W will be an ongoing trend for more accurate detection
°x» no matter in what ambient situation measurements
^important. Some research applications have a need to
50 Ppt to 100 ppt in order to measure accurately a
f No and/or N02 at concentration levels of 20 - 100
field test at a network station in the Black Forest
I^gg8yatem was used in parallel with a Luminox instrument
-&nd the NO trends were in good agreement but the
cr (by Unisearch Assoc., Inc., Canada) produced a
signal. This difference was in the order of 1-15J of
^-chamber instrument data .
User Interface
°an air P0lluti°n analyzers are microprocessor
and have RS 232 communication ports to be
e into a PC based data acquisition system. The
S are al80 totally controllable from the PC. All the
8 °f ^^e analyzer3 can be triggered and monitored
the Pc and Tecan provided software. If operated
m PC a11 functions are triggered from a simple
* y all nave a digital display for results and
An analog output is also provided.
e package is a programming tool. It allows the
» with the help of state of the art pull down menus
)JP Protocols consistent with his/her needs or legally
Procedures.
185
-------
Conclusion
This development project proves again that the
trying to find clues to explain phenomena in the environ*6
is best suited to work with instrumentation manufacturers
B*
V
J*
improve detection, signal and data processing technology- j
researcher is normally more creative in his approach than
manufacturer's engineers. But manufacturer's engineers
be more creative and consistent in producing a cost
effective, reliable and serviceable instrument.
The Tecan NO/NOx analyzer series 700 will be a signifi
-------
-^action
inte*fe*ences
time
Fi8ure
measure jterp measure
zero
S (measure-
mode)
NO
S (zero-mode)
dark-current
of PMT
TIME
ptgure
187
-------
HY
OF Ca AND Mg TO SELECTED
EASTERN KANSAS OAK-HICKORY
NG SEQUENTIAL SYNOPTIC EVENTS
Thomas Department of Civil
Engineering
University of Kansas
Lawrence, KS 66045
ar°tz Department of Civil
Engineering
University of Kansas
Lawrence, KS 66045
Lane Department of Civil
Engineering S
University of Kansas
Lawrence, KS 660*15
Th
!0r>t to Cnhivep3l-ty of Kansas is engaged in a three-year interdisciplinary
63 in ; lra°terize dry deposition at proximal forested and grass-covered
i 6ot (leaf f!Gotonal area. Both indirect (eddy flux and Bowen ratio) and
lrj°3''tion Washing) methods are being employed for assessment of dry
»e **8 the' A3 part of> the overall research design, data necessary to
,te' (b) w??'ture and degree of depositional variation (a) among individual
hin the Gan°Py of a tree, and (c) within and among sets of
Oondltloris (synoptic events) are collected. This paper
results from a portion of the 1987 sampling program.
k r*£ P* Q /
-nut- >S Oak> Quercu3 i"ubra; shagbark hickory, C_arya qyata;
Vlar Posif- ^-HSlans nigra) were selected for sampling. Leaves from
>ta?tlc ev °ns ln each sample tree were collected at the start of a
'Ho ^°.1 tf defined as the end of a previous period marked by rainfall
iU6ri^r>ati Leaves were washed and leaf areas were measured.
•ta] ^eaf of Ca and M8 were assessed by atomic absorption; mass per
at neci at ea was subsequently calculated. Ca concentration values
,H)D0oUl
-------
INTRODUCTION
Dr y _D^egO3 i_t ion
Various compounds are transported by the atmosphere and deposited or.
living organisms, the ground surface and other substrates. Many sue!-.
compounds have, or may have, detrimental effects on plant life, materials,
and ecosystems.
Airborne materials are deposited in several ways: 1) gaseous
constituents are adsorbed to surfaces or absorbed into various media,
including plant tissue; 2) rainout and washout processes carry and deposit
dissolved or particulate materials; and 3) aerosols attach to surfaces as a
result of dry deposition. The impact of aerosols on visibility and health,
their role in the transport of metals and certain organic chemicals, among
others, have received attention, but their effect in some areas of ecosyate:
research are unknown.
Moat research efforts intended to examine the influence of
atmospherically transported material on natural ecosystems have focused or,
rain chemistry and gaseous air pollutants. Recent evidence suggests that
significant quantities of material enter many ecosystems through dry
deposition. Results also indicate that the impact of dry deposition on
ecosystems may be quite important, especially since the process represents a
principal vector for the introduction of anthropogenic pollutants to natural
ecosystems (Lindberg et al., 1986; Graustein & Armstrong, 1983). Continued
work is clearly necessary, but examination of aerosol effects on terrestrial
ecosystems is hindered by a lack of: 1) sound methodology for measuring
aerosol inputs to ecosystems; 2) understanding regarding the interaction of
plant canopies with ambient air and aerosols; and 3) knowledge concerning
the chemical composition and size distribution of aerosols impinging upon
terrestrial ecosystems.
Current research on these three items is being done using a broad-
based, multi-element perspective (e.g., Lindberg et al., 1986), an important
approach if the magnitude and relative influence of different aerosol inputs
are to be determined. ft complimentary step involves focus on single
el-ernents and their interaction with the canopy in greater detail. From a
combination of broad and detailed studies, a generalized transport and
deposition model for an element can be constructed, and used to examine the
fate and effects of pollutant loads at a larger scale. The structural
outline of such an element model is shown in Figure 1 ; it involves
characterization of a) amounts and composition of dry deposition above and
below the canopy; b) precipitation amounts and chemistry above, within (on
leaves), and beneath the vegetative cover; c) the type, relative size, and
mass fractions of dry depositional products above, within, and below the
canopy; d) the relationships different depositional amounts have to weather
events; and e) the interactions of elements with selected trees, grasses and
forbs.
190
-------
Figure 1
AEROSOL-CANOPY INTERACTION MODEL
Atmosphere Subsystem
Plant Canopy Subsystem
I r~ ~ ~^~—~=. :r z—
Aerosol *H-| Synoptic Events ,
I
, Event '
l| M. J
I I J-
1 Depositional Products
Seasonal
Type
Dry
Wet
Amounts
Characteristics
I
I
I-
I
, H — — — — — ~ -i
_l_, Trees or Grass/Forbi
j1 Top of Canopy ^
11Within Canopy i
I1--! 1-1
lr~ — a—n f- — *~ — -~\
• iUptakei [Storage i
I
I
iBottom of Canopy
LU-------=F=-
Soil or Surface Water
An interdisciplinary group at the University of Kansas is conducting a
three-year EPA-sponsored project to characterize dry deposition at adjacent
forested and grass-covered sites. Various indirect and direct techniques
are being used to assess dry deposition within the framework presented in
Figure 1. Two goals of the project are to test and extend such models to a
variety of substrates under different conditions, and to assess
applicability within and among regions. The effort described in this paper
is representative of various current field projects intended to provide data
on single element charactristics for inclusion into a model structure.
PURPOSE
Most studies of directly-deposited material using leaf washing
techniques have given little attention to extant environmental conditions,
namely associations with large scale meteorological (synoptic) events
(Dasch, 1987). The purpose of this study is to assess the warm season,
event-based amount and rate of dry deposition of a single element onto
leaves selected from specific levels among three trees in a deciduous
forest. Events are defined here as periods between measurable on-site
precipitation, tree species selected are Quercus ruba, Juglans nigra, and
Ca£ya_ova^a, and total mass/concentrations per cm2 of calcium are measured
using a leaf washing technique. Calcium is used as the element of interest
because it is relatively stable on leaf surfaces (Lindberg, et al., 1986).
Two questions are addressed: 1) do depositional rates differ between two
similarly exposed but different height locations on a single tree; and 2) do
depositional rates differ among events on the same tree?
DESCRIPTION OF METHODS
Field Site
The field site lies within an oak-hickory-forest/tall-grass-prairie
ecotone in the Nelson Environmental Study Area approximately 15 kilometers
north of Lawrence, KS (90° 12' W and 3^° 03TN) at an elevation of 300 meters
above sea level (Figure 2).
The sample site is located a minimum of 1 kilometer from the southern
and western forest edges to insure that representative air flow across the
canopy exists (Figure 3). The site is as flat as possible to minimize slope
191
-------
effects on flow patterns, and is not located within 10 km of any major air
pollution sources. The forest canopy is 65 percent to 95 percent closed,
averaging 85 percent. Tree height is generally 16 to 18 meters; diameter at
breast height of most trees is between 30 and 45 cm. Common tree species
include Quer cus rubr a (red oak), Q^muhlenbergii (chestnut oak), and Q,
fnacrocarpa (bur oak), Carya ovata (shagbark hickory), Juglans niLg£a (blatf
walnut), Ulmus americana (American elm), and Fraxinus pennsylvarmja (greer,
ash).
Leaf Sampling
Leaf sampling was done within periods marked by similar synoptic
conditions, principally clear skies, average warm season temperatures ani
humidities, low speed wind flow from the south, and lower level stability.
Onset of a rain event marked the cessation of a sampling segment, and the
start of another. Three such segments spread over a thirty-day period
served as the sampling period (Figure 4). Period 1 was comprised of Julian
days 240-247; periods 2 and 3 extended from day 254 to 257 and 261 to 270,
respectively.
Leaves were collected shortly after the cessation of a rainfall event
in order to establish a baseline for subsequent measurements. All leaves
were taken from tree canopy sections with a similar aspect. Outer canopy
leaves from a height of approximately 13m were taken from the oak at two
positions; mid-canopy leaves were collected from the oak (8m), hickory
(10.4m) and walnut (10.4m). Ten leaves were gathered at each location on
individual trees; each leaf was handled with plastic gloves, placed in a
labeled, clean 250 ml plastic beaker, and capped'for transportation to the
laboratory.
Material Extrac^bioTi
Five blank samples were prepared by washing 250 ml containers with 50
ml of milli-Q water. Sets of five sample bottles containing leaves were
similarly washed (walnut leaves required 100 ml of water), and placed on
their sides on a rotating shaker for 3 minutes. Leaves were subsequently
removed with a forceps, blotted dry, labeled, and placed in storage bags for
refrigeration. The wash solution was brought to a pH of 2.0 using nitric
•acid.
Analysis
Analysis of calcium (Ca++) was done using a Perkin-Elmer 460 flame
atomic absorption spectrophotometer (AA). Standard concentration solutions
were made by diluting 100 mg/1 (1000 ppm) research grade calcium standard
with an appropriate amount of milli-Q water to obtain 10 mg/1, 1 mg/1, 0.6
mg/1, 0.4 mg/1, 0.2 mg/1, and 0.02 mg/1 concentrations. The AA was started
10 minutes before each analysis session and was aspirated with milli-Q water
when samples were not being analyzed. In order to avoid cross
contamination, the AA was aspirated with railli-Q water between sample runa,
and the sample tube was cleaned between samples. Standard concentrations
were run to produce regression curves. AA response over the concentration
range used (0.02 mg/1 to 1.0 mg/1) proved linear with a regression
coefficient of > 0.99.
192
-------
A 0.1 percent lanthanum solution was pipetted into all samples and
Agitated to ensure complete mixing. Raw data were recorded for each sample
using an averaging time of five seconds.
Leaf areas were determined using a Li-Cor Li-3000 portable leaf area
meter. Most leaves were measured as a unit, but some of the larger leaves
had to be cut into smaller pieces, especially the walnut leaves.
RESULTS
*a Deposition
Mean mass/cm2 of leaf area data for the oak samples are shown on Figure
4 for all three periods. All three samples show similar trends among
periods; deposition at mid-canopy was about 20 percent less than at one
apper canopy location for periods 1 and 2, but equaled or exceeded upper
canopy amounts in period 3. The suggestion is that within-canopy deposition
Is comparatively similar at the outer canopy throughout the mid-to-upper
levels in the oak at this particular location.
Values at mid-canopy level among the three sample trees were less
comparable (Figure 5). Deposition at mid-canopy on the hickory was about
ten to 300 percent higher than at the same level as the oak. The
depositional pattern in the walnut lagged that of the other two trees during
the first period, paralled that of the oak in period 2, and mirrored both
the oak and hickory in period 3. Ca also appeared to be retained better by
ualnut leaves, because precipitation did not wash as much Ca off the leaves
as was the case with the other two trees.
Leaf Areas
Mean values, of course, do not describe scatter in the composite data
at each sample location. Leaf areas varied substantially, and, thus,
affected the collection area available among individual trees. The upper
oanopy oak leaf area mean lay between 20 and 40 cm2 based on a comparison of
individual leaf area composite distributions for each sample; the data also
showed slight negative skew, and a relatively small size range. The scatter
in area values at mid-canopy in the oak was much greater than at upper
canopy levels; mean leaf size was larger (between 35 and 75 cm2), again with
negative skew, and was arrayed over a four-fold size range. Mid-canopy
hickory leaves showed composite mean values between 20 and 40 cm2 and a
symmetrical distribution; the size range was comparable to the oak samples.
Mnut leaves were the largest (mean values varied between 115 and 225 cm2)
and varied over a three-fold size range. The aerodynamic effects of such
disparate characteristics on collection efficiency at different locations in
the same tree and among trees are unknown. Tree geometry is, of course,
substantially different among the three trees also, and we are conducting
several laboratory simulations to address leaf size and arrangement
questions.
Ca Variability
Sample mass/unit area data show considerable scatter among periods,
«ithin the canopy, and among trees. Upper canopy oak deposltional amounts
showed a pattern of increasing scatter during period one, generally low
193
-------
variability at one of the oak locations in period two, but six-fol1
variation in amounts at the other oak location, and consistently larj;
ranges at both locations in period 3-
Mid-canopy values were generally more variable than those f rora t>.:
upper canopy. All three tree samples showed increasingly more disparat;
values as each period progressed. Oak samples showed the least overall
scatter; walnut values covered a range about twice that of the oak, whil*
hickory sample variability exceeded both of the latter by about a f actor c!
two.
Deposition and Meteorol ogical Events
Synoptic events within each period were characterized using National
Meteorological Center surface and 350 mb maps, and on-aite measurements of
precipitation, temperature, humidity, wind direction, wind speed, skj,
conditions, and pressure. All three periods began with frontal passage, an;
were marked by high pressure dominance, clear skies, low humidities, average
temperatures, southerly to westerly flow, mid-tropospheric subsidence, ar,i
slightly stable-neutral-slightly unstable conditions.
The degree of association among measured and derived meteorological
variables and depositional values was assessed by correlation analysis
(Figure 6). Deposition and meteorological variables were all in a manner
consistent with synoptic conditions associated. Significantly associated
values (0.05 level) in period 1 included temperature, stability and downward
motion. In period 2, only stability showed significant (0.05) association
with deposition; humidity, temperature and wind direction were significant
at 0.1. Finally, in period 3, temperature, humidity, wind direction and
stability were all positively and significantly associated with deposition,
In summary, periods marked by slightly increasing temperature and humidity,
low speed winds, direction favoring an easterly component, stability
conditions near neutral, and downward vertical motion in the lower troposphe
were associated with higher depositional amounts.
inpl ing Programs
We cannot make any definite statements about deposition and tree form,
or about depositional variation among different species because of sample
size. Additional trees will be sampled under the framework described so
that sample size does not prove limiting. We are also conducting wind
tunnel tests on leaf collection efficiencies among species; these data will
be helpful in estimating total bulk deposition.
REFERENCES
Dasch, J. 1987. Measurement of Dry Deposition to Surfaces in Deciduous and
Pine Canopies. Env. Pollut. 44: 261-277-
Graustein, W, and R. Armstrong. 1983. The Use of Strontium-37 /Strontium 86
Ratios to Measure Atmospheric Transport into Forested Watersheds. Science
219: 289-292.
Lindberg, S. , G. Lovett, D. Richter, and D. Johnson. 1986. Atmospheric
Deposition and Canopy Interactions of Major Ions in a Forest. Sc i enoe
231 : HJ1-146.
194
-------
KANSAS
JtHwton
Oouglu
'Gather Station
JEFFERSON COUNTY
&OUGLAS COUNTY
S$$J$$sNatui
^S^Re!
Sample Location
1mjla
Figure 2
T
1 kilometer
Site Location
195
-------
Weather
•-Sample Loco
lot* dot* from U.S.G.S. ropogroptiic mop,
Swin V«7t. Sf*«i 6962 II SW
•riginol icoU 1:24,000.
Figure 3 Vegetation I-Iap
196
-------
CALCIUM (Ca") DEPOSITIONAL AMOUNTS vs JULIAN DAY
OAK 1 OAK 2 AND OAK MID LOCATION
2.0
CO
OP
n>
-P-
o
fa
O
O
CO
§
rt
O
O
ss-
CO
O
<
g
w
o
a.
UJ
o
oOAK 1 LOCATION
+ OAK 2 LOCATION
xOAK MID LOCATION
I I I I I I I f I
240
245
250 255
JULIAN DAY
260
265
270
-------
CALCIUM (Ca~) DEPOSITIONAL AMOUNTS vs JULIAN DAY
OAK MID HICKORY AND WALNUT LOCATION
CO
00
cw
C
>-i
o
in
O
ro
13
o
V)
o
3
rt-
O
ex.
n
o
H-
0)
in
O
D»
D
ID
O
O
V)
o
Q.
U
Q
oOAK MID LOCATION
+ HICKORY LOCATION
xWALNUT LOCATION
240
245
250
255
260
265
270
-------
AEROSOL NITROGEN INPUTS TO A TREE/GRASS
ECOTONE: PROJECT OVERVIEW
Steven C. Mauch
Glen A. Marotz
Department of Civil
Engineering
University of Kansas
Lawrence, KS 66045
Department of Physics
University of Kansas
Lawrence, KS 660^5
Stephen J. Randtke Department of Civil
Engineering
University of Kansas
Lawrence, KS 660M5
Dennis D. Lane
Department of Civil
Engineering
University of Kansas
Lawrence, KS 660*15
Ray E. Carter, Jr. Department of Civil
Engineering
University of Kansas
Lawrence, KS 660^5
Mark J. Thomas
Peter G. Torrey
Department of Civil
Engineering
University of Kansas
Lawrence, KS 660M5
Department of Civil
Engineering
University of Kansas
Lawrence, KS 66045
Briefly described in this paper is an EPA sponsored, 3-year-
interdisciplinary study of nitrogen inputs to a tree-grass ecotone in
eastern Kansas being conducted by the Kansas University
Atmosphere/Ecosystem Interaction Research Laboratory.
The ultimate goal of the project is to model nitrogen deposition in an
ecotonal situation. The modelling strategy begins with ambient air
sampling and measurements of bulk, wet, and dry deposition to the
vegetative canopies. The focus will then shift to the character of
deposition to individual canopy elements during specific weather events.
Integration of these data into a generalized deposition model will allow
study of the fate of important anthropogenic pollutants in ecosystems.
In the first phase of the project, a detailed depositional data base is
being constructed for various constituents, with an emphasis on nitrogen.
Specific types of data currently being gathered include: quantitative and
qualitative aerosol data; qualitative precipitation samples; and background
meteorological data. The systems and protocols for collecting this data
have progressed from development to deployment. Meteorological, chemical,
and particle instruments have been tested in various combinations at both
research sites. Intensive sampling sessions are slated for the summer
field season.
2QO
-------
INTRODUCTION
Past research concerning atmospheric inputs to ecosystems has focussed
on rain chemistry and, more recently, on gaseous air pollutants. However,
a growing body of scientific literature suggests that the impact of dry
deposition on ecosystem structure and function may be quite significant.
In the context of this study, dry deposition is defined as the turbulence-
driven process of aerosol transport to surfaces in the absence of
hydrometeors. This excludes the consideration of dustfall, which involves
deposition of a particle fraction larger than true aerosols.
Dry deposition may serve as a major vector for the introduction of
anthropogenic pollutants to ecosystems (Lindberg et al.t 1986; Graustein &
Armstrong, 1983). Therefore, it is no longer sufficient to monitor only
the pollution loads present in bulk precipitation and gaseous form.
Atmospheric aerosols must also be monitored, and the interaction of these
aerosols with plant canopies must be studied if the effects of dry
deposition are to be understood.
A broad-based strategy is required to determine the atmospheric
depositional inputs to terrestrial communities, including systematic
measurements of aerosols and deposition both within and above the canopy.
Recent aerosol-canopy interaction research has begun to employ such
strategies, starting with the entire canopy as the basic unit of study
(e.g., Lindberg et al.t 1986). By narrowing the focus to include the
behavior of single canopy elements, a generalized model may be built for
transport and deposition of an element of interest at a landscape scale.
To succeed in building such a model, several hurdles must be overcome:
1) the lack of a sound methodology for measuring the aerosol inputs; 2) the
lack of understanding of plant canopy interaction with air masses and
aerosols; and 3) the lack of general knowledge concerning • aerosols in
"unpolluted" air. A research team at the University of Kansas has embarked
on a three-year, interdisciplinary project to address these problems. The
project's goal is the characterization of aerosol nitrogen inputs to a
tree/grass ecotone in northeastern Kansas. This paper is a brief overview
of two principal field research sites, the methodologies being employed,
and a brief presentation of some field experiences to date.
SITE DESCRIPTIONS
The research sites are situated within a natural preserve that is
devoid of significant proximal human-derived nitrogen sources. A 40 m
tower is located within a forested area at the north-central edge of the
reservation; a 10 m tower is located in an adjacent grass-covered area.
The forest tower is 1 km south of the prairie tower, and the two adjoining
areas are part of an oak-hickory-ash forest/tall-grass prairie ecotone.
The geographical location of the study areas is detailed in Figure 1.
The siting of the two towers was controlled largely by meteorological
fetch considerations. The prairie tower is situated in gently sloping
terrain, with an average cover height of 1 - 1.5 m during the growing
season. A tree stand east of the prairie tower limits adequate fetch
directions to a clockwise sweep from the east-southeast to the northeast.
In this sweep, the roughness lengths vary from 2 mm to the south-southeast
up to 12 mm to the north-northeast, with an average roughness length of
approximately 7 mm (as determined by wind profile measurements under
201
-------
neutral conditions). The forest tower is surrounded by at least 1 km of
continuous tree canopy in a 180° arc from east-southeast to west-northwest.
The forest canopy averages 85 percent closed, with trees at the site
generally 13 to 20 ra in height. Both sites have adequate fetch from the
prevailing wind directions, i.e. south and southwest.
In addition to the two research towers, a continuous recording weather
station has been established at the prairie site. The station includes a
10 m tower equipped to measure temperature, wind speed and direction,
precipitation, relative humidity, and insolation. The instruments are
monitored with an electronic data logger, and observations are* averaged
over ten minute periods. This station is designed to provide a long terra
climatological data base, as well as background meteorological data for
deposition experiments.
METHODOLOGIES
Aerosol Chemistry
The choice of a chemical sampling system was dictated by the need to
measure all chemical species of interest (particulate and gaseous) over as
short a time period as possible. The sampling system selected was the
glass-impactor/denuder-tube/filter-pack (GIDTFP) system (Stevens et al.,
1987). Each system consists of 1) a teflon-coated glass impactor, which
removes large particles (>2.5 \im aerodynamic diameter); 2) two annular
denuders, the first coated with Na2CC>3 to absorb acidic gases (S02, 803,
HNO^, and HN02), and the second coated with citric acid to absorb gaseous
fWj; and 3) a filter pack consisting of a teflon filter (to trap small
particles) and a nylon filter in series with a citric-acid-coated glass-
fiber filter (to capture HNC>3 and NHjj, respectively, volatilized from fine
particles). Air is drawn into the system through teflon tubes into the
glass impactors at a controlled flow rate of 16.7 LPM. The system is run
for four hours and can be used to accurately determine all the species of
interest. The two annular denuders in series reduce or eliminate various
artifacts associated with sampling nitrogenous aerosols by effectively
removing gaseous acids and bases.
The GIDTFP sampling system has been used by several researchers in the
past few years (e.g., Sickles et al., 1986, and Stevens et al., 1987), but
standard protocols for its use have not been developed, and questions
regarding its limitations remain. The first year of this study has
focussed on development of such protocols, and detailed laboratory study of
the glass impactors. The actual aerodynamic behavior of the impactors,
the presence (or absence) of large particles (>2.5 ym aerodynamic diameter)
bouncing past the impactor into the denuders, and possible artifacts
associated with compounds carried on large particles are being studied in
the laboratory. To facilitate examining these issues, our research team
conceived and designed a glass impactor with a removable impaction surface
mounted on the end of the denuder tube (see Figure 2).
Micrometeorology
The choice of appropriate dry deposition measurement techniques was
based on feasibility and comparability with previous studies. From the
wide array of methods suggested and used to quantitatively assess dry
deposition, those that fall under the category of flux parameterization
202
-------
best meet these criteria. Of the dozen or so flux approaches used in
various studies to this point, the Bowen ratio and gradient techniques
(Droppo and Hales, 1976; Garland, ]97^) were chosen.
The Bowen ratio method uses measurements of the surface energy budget
in conjunction with moisture gradients to derive the flux of water vapor
near the surface. Since the fluxes of any conservative scalar in the
atmosphere (e.g., heat, moisture, aerosols) are generally assumed to behave
similarly, aerosol fluxes (Fq) may be implied from the moisture fluxes
(Fw). The relationship used relates the measured moisture flux to measured
gradients:
- w2)
where Q^ , Q2 = aerosol concentrations at two levels, and
Wi , W2 = specific humidities at the same two levels.
A modified Bowen ratio method may be used when measurements of the flux
of moisture or heat are available from simultaneous eddy correlation
measurements. Eddy correlation will be discussed more with the gradient
method.
The instrumentation required for the Bowen ratio method consists of a
net radiometer, soil temperature probes, soil heat flux plates, and sets of
temperature and moisture sensors at two heights. A Bowen ratio system
developed by Campbell Scientific Co. is used; it uses fine-wire platinum
thermocouples to measure temperature, and a chilled-mirror dew point sensor
to measure moisture. The instrumentation is not capable of extremely fast
response times, but must be capable of high accuracy and precision to
measure small differentials that may be on the order of one percent. The
sampling periods involved may be relatively long; the budget terms are
computed for each 20 minute interval in an experiment. Bowen ratio
measurements are robust with respect to surface cover irregularities (fetch
conditions), making them applicable in a wide variety of locales.
The gradient method uses reasoning similar to that in the Bowen ratio
method, but with a different strategy to obtain the flux. Eddy correlation
is employed to measure the flux (F) of a scalar (q) in terms of a
diffusivity (K) , with
F = -K dq/dz
where dq/dz is the mean vertical gradient of the scalar of interest.
The value of K is determined by simultaneously measuring the variations
of vertical velocity (turbulence) and the reference scalar (e.g. moisture)
in addition to the mean vertical gradients of both scalars. The
measurements are performed over the time interval of interest at
frequencies on the order of 10 to 20 Hz, and the statistical covariation
of the time series is used to find K. K for the reference scalar is then
assumed to be identical to K for the scalar of interest. This method
requires fast response instrumentation that is also precise. A Campbell
Scientific sonic anemometer (model CA-27) is used to measure vertical
velocity, and an A.I.R., Inc. Lyman-alpha hygrometer (model AIR-LA-1 ) to
measure absolute humidity (moisture). Due to its responsiveness to the
aicro-scale turbulence structure, the gradient method is highly sensitive
to fetch conditions. This makes it less robust than the Bowen ratio
203
-------
method, although the two methods should be In good agreement where fetch i<
adequate.
FIELD EXPERIENCES
During the first year of the study, numerous field tests were
conducted, primarily for the purposes of developing protocols and verifying
methodologies. During the first year little attempt was made to integrate
the various aspects of the project, but integration will be the focus of
the second year activities. Selected field data follow; they are
representative of the types of data that will be gathered and integrated ir.
the second year at both the prairie and forest sites.
Chemical Data
Chemical data collected from three GIDTFP samplers mounted at different
heights on the prairie-site tower on 24 July 1987 (and run simultaneously
with the meteorological instruments) are presented in Table 1. Except for
gaseous S02f the concentrations of the constituents of interest were quite
low, many less than 1.0 yg/rn^. There was generally quite good agreement
among the three samplers, indicating that the sampling system can give
reasonably reproducible results. However, these data also illustrate some
of the difficulties encountered in trying to quantitatively measure
concentration gradients in rural ambient air:
1) Although the blank values are all less then 1 ng/m3, they are
significant relative to the concentrations found in the ambient
air. To quantitatively measure concentration gradients under these
conditions, the blank values must be as low and as reproducible as
possible. Otherwise, the analytical error associated with the
blank values will be very significant relative to the gradients
being measured.
2) Although similar results were obtained with each of the three
samplers, there is no evidence of a gradient for any of the species
analyzed. In fact, the results for the middle sampler are
generally higher than the results for both the top and bottom
samplers. Thus, the directions and magnitudes of any concentration
gradients that may have existed throughout the sampling period can
not be determined from these data. It is possible that there were
no gradients present, but it is also possible that the sampling
system did not give results that were precise enough to permit
gradients to be observed.
Several steps have been taken to minimize these difficulties during
subsequent sampling sessions:
1 . The extraction procedure was modified to give higher liquid
concentrations that could be more precisely quantified, and an
autosampler is now used to reduce analytical error.
2. Efforts were made to achieve lower and more consistent blank
values.
204
-------
3. Since the chemical results for the coarse particle fraction were
particularly erratic, an improved sampling inlet (Figure 2) was
developed to facilitate accurate sampling and analysis of the
coarse particle fraction.
4. Duplicate sampling systems are now being used at each of two
heights so that true differences in concentration with height can
be statistically verified.
On 13 November 1987, samples were collected using paired sampling
systems without glass impactors, while awaiting the construction of the
nodified inlets. Concentration gradients were measured for several
:onstituents, including S02, HN03, HN02, and sulfate, as shown in Table 2.
Unfortunately, meteorological data suitable for computing deposition
velocities were not available for that date. Nevertheless, the fact that
gradients were observed was very encouraging.
Table 3 demonstrates the difficulty encountered in attempting to
seasure chemical gradients in unusually clean air, in this case on a cold
day in January. Although the results were quite consistent for most of the
species analyzed, the concentrations of all species excluding fine
participate were below 1 ug/n)3 and only slightly higher than the blank
Samples were collected at the forest site on April 13, 1988, using the
nodified inlets fitted with oil-saturated 10-mm porous glass impactor
disks. Unfortunately, one of the impactor nozzles broke as it was being
attached to the denuder tube. The results for the three remaining sampling
systems are in good agreement (Table 4), and gradients in the
concentrations of S02 and sulfate were detected.
Micrometeorological Data
Two field tests using the gradient method with the GIDTFP sampling
system have been selected to illustrate the impacts of micrometeorology on
deposition measurements. The experiment at the prairie site on 24 July
1987 and the first experiment at the forest site on 13 April 1988 are
presented.
Prairie Experiment. — The prairie site experiment illustrates the
potential impacts of non-steady-state conditions. The sampling run was
conducted over a four hour period (0828 CDT to 1228 CDT). The instrument
array was positioned 1 m above the ground on the prairie tower, with a 1 m
separation between the GIDTFP systems and the temperature structure probes.
The sky conditions throughout the run were mostly clear, with scattered
small cumulus clouds appearing after 1030 CDT. A light dew was present at
the start of the experiment.
Time series of the five-minute means and standard deviations from the
fast-response instruments are presented in Figures 3a through 3e. The mean
temperature data show a strong warming, while mean temperature gradients
reveal the transition between morning stability and afternoon instability.
The five-minute covariance series of various scalars with vertical
velocity are presented in Figure 3f. The absolute humidity covariances
show extreme fluctuations relative to the other two covariances, which are
-ore consistent. This variation shows the impact of thermal turbulence and
evaporation of dew as conditions moved from stable to unstable. The
downward trend of the vertical velocity-horizontal velocity covariance may
205
-------
be explained by the change in wind direction from westerly flow at the
start of the experiment to more southerly flow at the end of the
experiment. This would be consistent with the lower roughness lengths
measured in southerly sectors, which illustrates an impact of variations ir.
fetch.
Examining the overall covariances relative to the block covariances
reveals another effect of non-steady-state conditions. The covariances
taken over the entire run are denoted by the symbols on the right margin of
Figure 3f. The overall mean covariance of vertical velocity and
temperature is negative, whereas all the block covariances are positive.
Lower mean temperatures at the start of the experiment were associated with
higher mean vertical velocities, creating a negative bias in the
covariance. Therefore, fluxes calculated using heat as the reference
scalar could indicate suspension rather than deposition (if moisture had
been used). Such a reversal is contrary to the assumption of similarity
that is used in the gradient method.
Forest Experiment. — The first experiment at the forest site was
conducted on 13 April 1988. This was a 3.5 hour run, due to generator
problems. The run began at 1215 CDT and ended at 1615 CDT, with 30 minutes
down for repairs from 1330 to 1400. The weather conditions during the run
were marked by a cloudless sky and light southwest winds. The instruments
were stationed at 16 m on the forest tower, just above the leafless canopy.
The separation of the GIDTFP systems and temperature structure probes was
1.5 m. An electronics failure in the Lyman-alpha hygrometer made fast-
response moisture data unavailable for the entire experiment. The one-
minute means from the fast-response instruments and the one-minute
covariance traces from the forest experiment are presented in Figures i)a
through 4f.
In contrast to the prairie experiment, the micrometeo'rological
conditions during the forest experiment were quite consistent.
Unfortunately, no evaluation of the validity of the similarity assumption
is possible without the moisture data. The overall covariances are
consistent with the block covariances for this data set. By combining this
meteorological data with the chemical data, a deposition velocity of
approximately 9 cm/sec was calculated for gaseous sulfate. This deposition
velocity is somewhat higher than typically reported values for sulfate,
but can be considered reasonable within the limits of experimental error.
CONCLUSIONS
The results from forest and prairie experiments presented here
illustrate the hurdles that stand in the way of a greater understanding of
aerosol nitrogen inputs to a tree/grass ecotone. The first year of the
project has focussed on the acquisition of instrumentation, testing the
instrumentation, and developing the protocols necessary to overcome these
hurdles. To date, the following specific milestones have been
accomplished:
1) Establishment of a base meteorological station to provide
continuous long-term data to complement the intermittent finer
resolution experimental data;
206
-------
2) Construction and instrumentation of two research towers to measure
dry deposition velocities with the gradient and Bowen ratio
methods;
3) Refinement and improvement of the GIDTFP system based on field and
laboratory testing;
1) Development of specific protocols for sampling and chemical
analysis of aerosols at low ambient concentrations;
5) Development of software to handle the large quantities of
meteorological and chemical data generated by field tests;
6) Demonstration of the feasibility of integrating the above
equipment and protocols to characterize nitrogen deposition under a
variety of field conditions.
These experiences to date show that the future research thrust should
De in the following areas. First, a more complete characterization of the
jIDTFP sampling system is needed to define the systematic limitations on
Jry deposition monitoring. Second, criteria must be developed to allow
fast-response micrometeorological data to be interpreted with the longer
time scale GIDTFP data. Third, more accurate methods need to be developed
to quantify near-surface ambient concentration gradients in rural air.
The summer of 1988 will be devoted to intensive field work, especially
at the forest site. Dry deposition work will be supplemented by a study of
precipitation and throughfall chemistry at the forest site. This will more
completely define the atmospheric inputs of nitrogen and other compounds to
the ecosystem. The results from this field season will build a substantial
data base, providing results that should find broader applications in
related areas.
207
-------
REFERENCES
PP , . an Hales, 1974. Profile methods of Drv Dpnn^iHnn
*!p?^emeRnt; Im En*elma™> R. and G. Sehemel (eds.) Atmosphere--
Surface Exchange of Particulate and Gaseous Pollutants. Washington
D.U, Energy Research and Development Administration, p. 192-211.
Garland, J., 1974. Dry Deposition of S02 and Other Gases, in:
P^M1^?' R'H annd G' Seherael (eds-> Atmosphere—Surface' Exchange of
Particulate and Gaseous Pollutant... Washington, B.C., Energy Resear
and Development Administration, p 212-227 ^nergy ftesearc,.
'
paper submitted for publication (1987).
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support provided for this
'Th *«,"? U'S- Envir°™-fcal Protection Agency
The KU Experimental and Applied
208
-------
a le 1. Chemical Data for the July 24, 1987, Prairie-Site Samples
5 ^e art. so,
p ne Part. NO
P n6 ?art- NO
p e Part- NH
P . Mas3
. SO,
's
— -. NOj
Part. NH,,
Top
15.-48
0.65
O.J48
0.34
3-21
0.4?
0.14
1.09
22.5
3.51
1.41
0.25
1.09
Concentration, yg/m3*
Middle
16.85
0.76
0.46
0.57
3-49
08
1
0.1
1.1
50.0
4.17
2.19
0.23
1.13
Bottom
14.79
0.68
0.48
0.38
2.97
0.66
0.14
1 .05
17.5
3.23
1.55
0.27
1.05
Blank Average
0.03
0.03
0.00
0.00
0.17
0.42
0.39
0.26
0.35
0.47
0.43
0.40
15.71
0.70
0.47
0.43
3.22
0.74
0.14
1.09
30.0
3.64
1.72
0.25
1.09
corrections have been made to the sampler data; the average:
ape the averages for the three samplers.
6 2:
Chemical Data for the November 13, 1987 Prairie-Site Samples
Noa
NO,
Concentration, pg/tn3*
Top #2 Bot./M Bot.*2
Top Bottom Overall
Blank Average Average Average
~"
4.62
0.73
0.30
0.80
4.38
0.78
0.27
0.82
2.92
0.65
0.05
0.60
2.86
0.42
0.22
1.08
0.16
0.03
0.00
0.05
4.50
0.76
0.29
0.81
2.89
0.54
0.14
0.84
3.70
0.65
0.21
0.83
0.66
0.31
0.35
0.08
o.o
0.78
0.56
1.60
0.08
14.3
0.53
0.37
0.40
0.10
0.0
0.40
0.59
0.78
0.08
20.0
0.12
0.07
O.H8
0.10
0.0
0.72
0.44
0.98
0.08
7.2
0.33
0.22
O.J44
0,10
10.0
0.52
0.33
0.71
0.09
8.6
have been applied to the sampler data.
209
-------
Table 3. Chemical Data for the January 22, 1988 Prairie-Site Samples
Con centra t Ion , ,
**
Constituent Top #1 Top #2 Bo t ._# ]_ _ Bot.#2 Blank
S02 (as SO,
HN03
HNOZ
NH3
Part. SO,,
Part. N03
Part. NOZ
Part. NH4
Part. Mass
0.63
0.06
0.00
0.20
0.50
0.00
0.00
0.15
0.50
0.00
0.02
0.15
0.50
0.00
0.00
0.15
0.25
0.12
0.12
0.09
0.57
0.03
0.00
0.18
0.50
0.00
0.01
0.15
0.25
0.05
0.00
0.08
98.3
0.28
0.19
0.39
0.08
29.3
0.29
0.03
0.00
0,08
49.0
0.29
0.03
0.00
0.08
31 .0
0.1 1
0.43
0.53
0.16
0.0
Top Bottom Overa','.
Average Average Averag;
0.53
0.02.
0,01
0.16
0.24
0.08
0.10
0.08
51.9
• 0.27
0.12
0.20
0.08
63.8
* Blank corrections have been made to the sampler data.
0.29
0.03
0.00
0.08
40.0
Table 4. Chemical Data for the April 13, 1988 Forest-Tower Samples
Concentration, \ig/m3*
Top Bottom Overall
Constituent Top_JM__ Top _#2 Bot. #1 _ Bot.//2 Blank Average Average Average
S02 (as
HN03
HN02
NH3
FP SO,,**
FP NO 3
FP N02
FP NH^
FP Mass
TP S04
TP NO 3
TP N02
TP NHU
6.48
3.12
0.30
1.15
1.64
1.76
0.09
0.88
3^.3
2.17
2.4?
0.16
1.16
6.49
3.21
0.31
1 .22
1 .66
1 .81
0,16
0.64
65.7
2.42
2.64
0.26
0.74
4.95
3.19
0.34
1.32
1.54
1 .84
0.16
0.75
41 .7
1 .70
2.59
0.28
1 .03
0.77
0.09
0.08
0.10
0.21
0.48
0.50
0.50
ND
0.55
0.83
0.57
0.68
6.49
3.17
0.31
1.19
1 .65
1 .79
0.13
0.76
50.0
2.30
2.56
0.21
0.95
4.95
3-19
0.34
1.32
1.54
1.84
0.16
0.75
41 .7
1 .70
2.59
0.28
1 .03
5.72
3.18
0.33
1.26
1.60
1.82
0.15
0.76
45.9
2.00
2,58
0.25
0.99
* Blank corrections have been made to the sampler data. The glass impaetcr
for bottom sampler #2 broke while being connected to the denuder.
** FP = fine particulate; TP = total participate
210
-------
T
11
S
32
JEFFERSON COUNTY
T
12
S
•Nelson Environmental!.Study Area'/
|Rockefellerg
^Experimental
£ Tract i||
33-v
34
SB
Weather Station
g-Sample Location
1 mile
0
Figure 2
1 kilometer
Site Location
195
-------
IU
ro
_k
CO
(a)
5-minute Moan Wind Velocities (1 m)
I
(c)
5—minute Mean Temperature Gradients
E
V
(b)
Figure 3. Meteorology for the 24 July 1987 prairie experiment
IU
-------
5-minute Mean Absolute Humidities
Mean 1-minute Wind Velocities
Forut EjcfMrimmt on 4/13/68
ro
5—minute Mean Covariances
7/24/B7 by 1
(f)
Figure 3. (cont'd)
N
E
(a)
Mean 1-minute Vertical Velocities
FofW Eupvimtnt on «/'V««
Figure 4. Meteorology for .13 Apr±l 1988
-------
N
o
E
ro
__i
01
16O I BO
(c)
1—minute Mean (W,U) Covariances
Forwt Experiment an +/' 3/88
(d)
Winutec into run
.,.. (e)
Mean 1—minute Air Temperatures
L on 4/13/B8
kGnute* into nift
(f)
Figure 4. (cont'd)
-------
TOXICS IN FOG AND THEIR POTENTIAL
ENVIRCNMEWTAL PROBLEMS
Kumar Ganesan, Ph.D.
Dept. of Environmental Engineering
Montana College of Mineral
Science and Technology
Butte, Montana 59701
Research into the character and origins of acid rain has led
scientists to discover that clouds and fog are much more acidic
precipitation itself.
Recent studies identified several contaminants in fog droplets. Sere
of the chemicals identified are: volatile organic acids (VGA),
dicarboxylic acid (DCA), aldehydes, alkyl sulfonates, pesticides, trace
metals in addition to the low pH (^2), sulfates, and nitrates. Studies
indicate that the fog droplets act as an excellent scavenger of many
gaseous and participate pollutants, such as HN03, H202, ^3, i H20,
lulfates, nilrates, and suspended particulates. Sore of the pesticides
identified in fog droplets were over 3000 times higher in concentration
than could be accounted for by Henry's Law.
Fog consists of droplets, 30-40% of which are in the respirable
range Direct inhalation of such droplets or indirect contact could
increase the risk of individuals exposed to such contaminants, especially
in areas where identifiable toxic sources occur together with high
frequent fog and atmospheric stagnation. Since fog droplets have a much
hiqher deposition velocity than the contaminant themselves. This provides
a much higher chance for the contaminants to deposit on sensitive forest,
crop, soil and water bodies compared to direct deposition. Specifically a
fog/cloud could be an effective carrier of trace contaminants including
H202 to sensitive high altitude forests and cause potential damage.
Immediate research needs are also discussed.
216
-------
' ^reduction
into the character and origins of acid rain has led
Pitafft° d*scover that clouds and fog are much more acidic than
jg rT.j1011 itself. Though questions about such low pH observed in
informs • are yet: to ke resolved, characterization studies have added
rmation reoardina the tvoes of chemicals nresent in foa and
regarding the types of chemicals present in fog and
Organic acids(!•2), aldehydes(3), alkyl sulphonates <4'5),
have hoi ^^ trace metals(7) are a few of the chemical substances
?**ventio i6" *dentified in the fog and/or clouds in addition to the
!?U3 are tt^I. inor9anic sulfur and nitrogen compounds f8'9). Fog and clouds
only very acidic in nature but are effective carrier of
"eluding toxics. The potential impact of fog and clouds on
rstand"1' ^ama9e to crops and materials needs to be evaluated to
potent•and.assess the risk involved. It is also important to analyze
tial impact on natural resources such as forest, water and soil.
d?als witn characteristics of fog, physiochemical
f°9 on Plants and forests, deposition
Toxics in fog,
scus t
ai *r Potent-- Gd "^asurements and techniques of sampling.
80 Disc env^ronmental problems and remaining research needs are
1 - , . , -..-..- .
n ?ents also have been reported in fog samples. Fog being
anrt ssiPated at the ground level and having the potential to
f°9s cWT ^ransPort excess levels of chemicals, human exposure to
hlricallvInin9 toxics could cause adverse health effects.
n almost all air pollution episodes have been associated with
°^ f?fi Professor J- Pirket at the University of Liege
to 5th of December, 1930, a thick fog covered a
°? ^iQ11111* along the Meuse Valley. A large number
injured, several hundred were severely attacked
troubles, and 63 died on the 4th and 5th of
the 6tn of ^^ecember, the fog disappeared;
troul°les improved. Wherever fogs of several days
^rec3uent' public authorities were anxious to know
e fcll^s catastrophe. This apprehension was quite
be fn' Proportionally the public services of London
Dhpn Ced with tne responsibility of 3200 sudden deaths if
ncnenon occurred there."
*3 ntween 1873 and 1892 about 2605 excess deaths, and between
745° excess deaths were reported due to air pollution
these episodes were associated with the occurrence of
fa13*161106 of the Problem is "e11 established; but the
, and its specific causes have never been adequately
1.1 Reasons For Concern
in ambient air and tnose directly emitted from
s s ln urban and rural areas may be effectively scavenged by
areas Uhh fogs may ** of c00061" to public health in specific
fog Jf1 er? t*le occurrence of fog is frequent. Enrichment of
' that is, more vapor dissolution in fog droplets than would
217
-------
dissolve in an ideal solution at equilibrium, has been reported for
pesticides. For example, an enrichment of about 3200 was reported for
pendiinethalin in fog water samples in Parlier, California^).
Because of fogs low pH<8), deposition from fog and clouds on
vegetation crops and forest canopies have been the focus of sane recent
studies'13"15). Vegetation can act as an effective collectors of fog
droplets. Thus vegetation could be exposed to relatively high
concentration of chemical species by the deposition of fog droplets thar.
by simple dry deposition, direct exposure or even precipitation. Because
fog droplets have a much higher deposition velocity than toxics by
themselves as gas or particles, the scavenging of toxics in fog gives
these toxics a much higher 'chance to deposit on sensitive receptors. This
problem has not been addressed so far in the literature. The possibility
of such trace toxic compounds in high altitude fogs and clouds has also
not been investigated to evaluate its added effect, if any, on the alreai
stressed high altitude forest system.
Due to its high deposition velocity, fog may be a potential source tf
chemical input to lakes and soils. Chemicals have a much higher chanced
being deposited into lakes and soils when combined with fog droplets than
by direct deposition.
In addition, damage to materials exposed to fogs containing chemicals
has already been reported'12). The impact on materials damage due to
trace substances in fog, especially in frequently foggy areas, needs tob
evaluated.
Sources of pollution such as hazardous waste dumps, incinerators,
wood burning, pesticide spraying, combustion of fuel, and industrial
effluents have been shown to emit an array of toxics and trace metals.
These sources are in many cases located in thickly populated, foggy areas.
The impact of pollutant-laden fog becomes a concern in such areas.
2. Characteristics Of Flog
Fog is a ground level cloud formed by cooling of moist air below its
condensation point. Based on their different cooling mechanisms involve}
fogs are considered either radiation fog or advective fog.
On clear nights, radiative cooling of the ground due to net upward
heat flux causes the air close to the ground to cool. Cooling of the
moist air sets the stage for radiation fog. As cooling progresses, the
fog thickens, increases in depth and persists through the night. After
dawn, the net downward flux of heat from the sun to the ground dissipates
the fog. However, a thick fog can effectively reduce the incoming heat
flux, resulting in prolonged fog periods. Coupled with light wind
conditions, such thick fog may persist for days.
In contrast, advection of large scale warm moist air onto cool land
mass can establish intense thermal gradients in the lower air mass leadin
to fog formation. This is advective fog. Eddy mixing also is essential
for fog formation, however, a more violent mixing can inhibit the
formation of fog.
2.1 Frequency Of Occurrence Of Fog
Hardwick compiled fog statistics from 244 first order weather
stations in the United States'16). These data present annual mean numbei
218
-------
important observation by Glotfelty et al., ia the high
factor for pesticides in fog water, which is far in excess of
ns accounte(i for ty Henry's Law(5>. There is no clear
that the" E°r this observed enrichment. However, Glotfelty postulates
in the fj?re?ence of non-pesticide soapy and foamy organic matter observed
"ht act a a thin-film brrir r
the esr "^ht act as a thin-film barrier around the droplet preventing
iS? t^PP6*31 pollutants. The implication of the enhancement
there are possibly several other toxics which may behave in
1 and exist in higher than expected concentration in fog.
tn*3 enhancement, fog behaves as an effective scavenger of
especially toxics which may be of concern to health.
esS^idual lifftime r*sk for a single compound at one TSDP facility
s that at 10~5 P81" yeaim "Ihere was n° mention about the superfund
i are un<3ergoing clean-up. This may not be a permanent sources;
areas may be affected during the clean-up process.
'3 Impact Of fog On Plants And Forests
?wisa cha*!* et al-^38'r reported spot-lesions and extensive pockmarks on
b09 Anoft?! and table beets when exposed to fog on November 22, 1959 in
S04. oil; 1** acidity of the fog was reported to be 144 yg/ta3 as
of yellow birch misted with sulfuric acid at pH of 2.8
spots after one or more exposures. Similarly an
Caching of Ca2+, K+ and amino acids was also
: A wide variety of plant species reported to have shown
injury when misted with droplets in the pH range of 1.7 to
8 3t ^^?h elevations have been observed to experience injury to
<3eclinin9 growth in Europe as well as in North America. High
^-^P16^ w^th lengthy periods of cloud immersion and coniferous
on S°ntrib^te to potentially high cloud droplet capture.
, tr of forests in clouds carrying high concentrations of mineral
ace metals and ^^ c5"1 ^e a source of water and chemical
Lovet^' €t al.^41), measured and modeled such deposits in
am fir forest in the inountains of New Hampshire. Annual
vuc lti-0n by cloud droplet capture and bulk precipitation was
Unt -1 elevation of 1220 m in a 10.3 m tall balsam fir stand on
Hampshire. The results indicated that cloud
Ha + alone contributes to 60 to 80% of the annual deposition of H+,
' K , SO2", and NO"1 ions as shown in Table III.
4 3
measured the occult precipitation which includes fog,
at Great ^ Fel1' Cumbria in the United Kingdom. Flux
VSQ re taken bv niicroroeteorological techniques in an area with
surfaS on consisting of short grass roughened by areas of exposed
ts Q rocks and moorland plants. A flux of 20 mg/m2/sec of cloud
to i tile vegetation was reported. They concluded that in areas
0j!' cl°uds sampling would underestimate the total chemical
y 20% if occult precipitation is not included. The magnitude
of the cloud contact with the forest canopy is a function
topography, type of forest and the frequency of
a ground level cloud. The frequency of occurrence of fog is
219
-------
4.1 Sources^Of Toxics. Thomsons et al.(35), analyzed the air
toxics problem in the United States and assessed the cancer risk posed by
selected pollutants. They concluded that a wide variety of sources
contribute to individual risk and aggregate incidence of cancer from air
toxics. These include: road vehicles; combustion of coal and oil; wood
stoves; metallurgical industries; chemical production and manufacturing?
gasoline marketing; solvent usage; and waste oil disposal. They also
reported a list of pollutants that may be important contributors to
aggregate cancer incidence. They are: chromium, arsenic, asbestos,
products of incomplete combustion, formaldehyde, benzene, ethylene oxide,
gasoline vapors, chloroform, carbontetrachloride, perchloroethylene and
trichloroethylene.
4.2 Impact^ftnd Health Concerns. Literature on the health effects
of fog is very minimum. Deal discussed the potential adverse health
effects due to the acidity of fog particles, but concludes that H+ ion
concentrations are not high enough to pose a health problem(36)t However,
he does not consider fog particles that are trapped in the upper airways
and the possibility of their constituents being transported to the lungs
has not been evaluated for their adverse health effects.
Silbergeldf37) indicated that cancer is not one of the more
widespread toxics-induced diseases, although it is serious when it occurs.
Silbergeld was concerned that exposure to toxic pollutants may adversely
affect the nervous system and the immune system, causing intermittent but
frequent decrements in function. Recent research findings at the
University of Maryland by Broadwall and colleagues demonstrated that there
is no physiological barrier to prevent inhaled smoke from being directly
taken up by the brain. Large molecules like lead, when placed in the nose
of primates, seem to be transported to the brain in a retrograde fashion,
Thus there appears to be no nose-brain barrier. This specific finding is
of great concern because even with the effective removal of particles of
greater than 10 m diameter by the respiratory defensive mechanism those
droplets can still contain toxic pollutants which may possibly be
transported to or directly taken up by the brain from the nasal passages.
Trace metals and toxic substances that are carried by the droplets
have not been discussed in the literature. For example, lead was found in
fog in concentrations of 2540 yg/1 in Pasadena, CA, during November 15,
1981. Iron, manganese, copper and nickel are the other trace metals
measured in fog.
Recently, Glotfelty, et al.(6), identified several pesticides in fog
samples in Beltsville, Maryland and the San Joaquin Valley in California.
Aerosol particles and nonpesticide extractable organic matter was found in
all the fog samples. The Beltsville samples were much more acidic
(pH ^2.42) than the California samples (pH ^ 5.1-7.0}. They also
identified organophosphorus insecticides and herbicides in the fog. The
California fog samples showed higher concentration and greater variety of
pesticides than Beltsville fog samples as shown in Table II.
The California fog samples contained a variety of toxic oxygen
analogues of organophosphorus insecticides. Parathion oxygen analogue
(the major component of the axons) concentration in the Lodi, CA site was
as high as 184ug/l and parathion was 51.4yg/1, making it the highest
observed source in California. It should be noted that axons can be
formed in gas-phase reactions of the parent insecticides with ozone.
Axons are potent choline esterase inhibitors and are responsible for the
toxic effects of the organophosphorus insecticides.
220
-------
or ' especially formaldehyde concentrations, in Linnex,
it raranged frcm 460° to 12'800 M 9/1'* and in Oildale, Bakersfield
^ 610° to 1440° V9/1- Aldehydes are emitted directly from
Sources and also formed in the atmosphere by photochemical
U g/i .' ^e presence of other aldehydes such as acetaldehyde up to 170
conf •acetone/ acrolein, propanol combined together up to 840 y g/1
are rai?d by Grosjean and Wright (1983) in fog samples. Very little
^^ilable to compare these measurements. Zafinon et al.l25)r in
of aicie^S11 Islands, considered as a remote location, reported 0.008 mg/1
^ °°ntin *n tne ra^n water' Measurements in Ireland and West Germany
rainwater?26ally influenced locations showed 0.140 mg/1 of aldehyde in
*ct Qs '• It appears that hydrcmeteors, especially fog and clouds,
17 ^3S P*1336 carbonyl scavengers. Particle scavenging does not
carbonyl scavenging because particulate-phase concentration
a- "to? several orders of magnitude lower than the gas phase
• onst ' 8) However, formation and removal of carbonyls by free
droplets should also be considered (29f30).
et al.(4), identified hydroxymethane sulfonate ion (HMSA) in
cloud water. Concentrations up to 300 micromoles per liter
^a51^6*3 in Bakersfield, California within 5 kilometers of
°f sulfur dioxide. The formation of HMSA in fog water
05*8 an excess of S(IV) and CH20 in the droplet phase in
his forma*- the H60^'3 law equilibrium. Because HMSA is a strong acid,
h^ formai?n can te significant in the acidity of the fog. In addition,
°2 and n f^i18-!3.130 allow the coexistence of S{IV) and oxidants such as
oentrat? } • Jacob, et al . ( 7 ) , Munger et al . ( 5 ) , measured
Hea i^J113 far in excess of Henry's law equilibrium in fog water
20. *:" San Joaquin Valley, up to 3 x 10~3M of S(IV) and 7 x 10"4M
8 reporfSParent iinearity between HMSA concentration and CH20 and S(IV)
f^' howl Hi9h S02 and intermediate pH seem to favor the formation of
ePresentaVer'.Preservation of HMSA requires lower pH. Olius HMSA
s an important source of acidity in the droplets.
to! tain to?ay' et al.(33), reported (Ci-Cxo) volatile organic acids (VQA)
ihter and • in New York' Kawamura and Kaplan(2), measured VQA in rain
i? Los ^Q1!} f°9 in Los Angeles. Concentration of VGA (138.4yM) in fog
rM y u? es .was about six times higher than that measured in rain water
cv-rlc acids (cr^3 acids). Kawamura et al.t1), measured
204) c acid in fog up to 65 M in fog in Los Angeles. Oxalic
repoSJ?Ccinict (C4H504) acids were the most abundant dicarboxylic
o monitored two fog(34) events in the Po Valley,
total rmi-ne f°9's scavenging efficiency. Sampling for major ions
* showSSpended particulate matter (TSP) before and after each fog
luwed a ran of 33% to 79% removal of -rgp UH+ N0- a^ S02-. I
434
of f°9 water for typical polluted atmosphere with air
n °f 3, 5, and 1 ppb respectively of HN03, NH3 and H202 Jacob
r 100% scavenin of ^0 • ^ and H0 in 30 minutes
f°9 droplets act as an excellent vehicle for several fine
and gaseous compounds.
4.0
Toxics 'in Pog: Sources, Their Impact And Health Concern
221
-------
of days with heavy fog. Heavy fog is defined as fog capable of reducing
visibility to less than I/4 raile { 400m). The East and West Coast and
the Southeast Coast experience heavy fog more often than the rest of the
country. In Wisconsin, areas around lakes, fog occurrence also is
relatively high. Pennsylvania appears to have visibility less than 1/4
mile for more than 70 days per year. This annual picture does not indsl
days that have fog conditions but the visibility is more than 1/4 mile.
Thus, it is reasonable to expect that actual fog occurrence will be hiqte
with marked local variations in frequency. It should also be noted that
the areas where fog occurrence is high, are densely populated.
2.2 Particle Size Distribution Of Pog
Fog consists of aqueous particles ranging from 2 to 200 micrometers
in diameter. Cwensf^J, Ludwig and Robinson(I8) and Junge^^) studied
particle size distribution of fog and clouds. Ludwig and Robinson
reported particle distribution of fog for four conditions: in cloud,
below cloud, lower edge of cloud, and during cloud dissipation. A
continual increase in the number of droplets with decreasing diameter was
reported. Maximum number of particles were observed in the range of 4 to
16 micrometer diameter. Mack et al,'20)/ obtained haze and fog size
distribution at 13 and 42 m above ground near Los Angeles and Vandenberg
Air Force Base (AEB) in California, prior to, during and after fog events,
Los Angeles fog was characterized by higher liquid water content (DHC) an;
smaller size drops. About 30 to 40 percent of the droplets were in the
respirable range of 5-6 micrometer in radius.
Rodger et al.(21), measured particle size in several advection fogs
at Vandenberg AFB and reported a range of 3 to 40 micrometer in diameter
with a mode of 8 to 12 micrometers. Rodger et al.(22), with additional
measurements reported a peaking of droplet spectra at diameters 8 to 16
micrometers. Dickson et al,("), measured four fog events near
Capistrano, California, using a forward scattering laser fog nephlcmeter,
The concentration of droplets increased as droplet size decreased to abou
15 micrometer diameter. It shows that more numbers of smaller size
droplets exist in fog.
Thus, it is reasonable to assume that about 30 to 40 percent of the
fog droplets exists in the respirable range size.
3.0 Chemical Composition Of Pog
Jacob and Hoffman^24) studied the chemistry of fog formation using a
hybrid kinetic and equilibrium model. Their input data were based on the
concentration measured in an air mass in Los Angeles prior to fog
formation at a polluted site. The model results showed 97% and 100%
scavenging of NH3 and HN03 by fog in 30 minutes. The aqueous phase
oxidation of S (IV) was a maj, source of sulfate formation in the drople
with the oxidants 1^02 and 02 catalyzed by Fe (III) and Mn (II). However
if the pH >5 then ozone becomes important in oxidation process. Fog does
not seem to affect the S02 gas-phase concentration significantly.
Fog water collected in Los Angeles and San Joaquin Valley in
California usually indicated high concentration of N03~, SQ^, NH4+ and
HT1". The pH of the fog samples varied from 2.2 to 5.7. In addition, tra<
metals like lead were measured up to a maximum of 2540 yg/1 in Pasadena
fog during November 15, 1981<9).
222
-------
stafthe coastal region* east, west and southeast coasts of the
ncp u and *n re9i°ns around lakes. For example, Pennsylvania
encp u . ,
than ^n avy f°9 conditions, with visibility less than 1/4 mile, for
«n 70 days in a year.
°0l^6r2t^0ntains water droplets that are formed by water vapor
Various st°n °n nuclei fay heterogeneous processes. The droplets are of
S with a mode of 8 to 12 micrometer in diameter.
30-40% of the droplets are in the respirable size range.
"^emcr ' chanical characterization studies of fog, focused mainly in
foq *:lng the acid precipitation chemistry, leads to the revelation
ch^i aroplets are low in pH and contain several contaminants. Some of
%des i ldentified are: volatile organic acids, dicarboxylic acid,
88(1!? i yl sulfonates, pesticides, trace metals in addition to the
v 4), sulfates and nitrates.
Q j §
S^ 9asei^S indicate that f°g droplets act as an excellent scavenger of
kl°* of tii particulate pollutants. For example, fog could scavenge
Sf°re and6 2N°3/ H2°2 frcm the ambient air in 30 minutes. Field sampling
roent rL^ ter one f°g event in Po Valley, Italy showed 64, 71, 87
«noval of secondary particulates of SO2", NH+ and NO"
resPecti 44 3
vely and up to 79% removal of total suspended particulates (TSP).
Sciflg f.f ,.
hiohe • e Pesticides identified in fog droplets were over 3000
1 ?":r ln concentration than could be accounted by Henry's Law.
be Jr? roetals are identified in fog droplets. Thus, fog droplets
**n important sink for many gases and particulate trace compounds,
.toxic substances. In other words, fog could be an effective
, ^a transporter of such trace compounds. Inhalation of such
indirect contact could increase the risk of people to such
'• In populated areas, with both frequent fogs and frequent
Oo~ristagnation conditions, fog could be a potential source of
^rn in the event of any toxic sources.
Ue to it"6 ^"a^e to forest and crops by acidic fog is well documented
« ^"5— w w^ i».u •» v«* *%A *-•-*. ^o^h^ ^J *-*>—•J-'-*^-*— J-V^H <^<9 WC?^^ UV^wULU&l 1W^.^*1
hiah ^ and deposition of chemicals. Because fog droplets have
• " er deposition velocity than the toxics by themselves as gases
the scavenging of toxics and pollutants in fog, gives these
a much higher chance of being deposited on sensitive
' such as forests, crops, water bodies, and soils.
Possibility that toxics may be transported by clouds and fog to
ressed forest ecosystems, especially in high altitude areas, has
evaluated thus far.
•Qg acti^6 P°tential environmental problems with fog can be summarized
a major sink for toxics, gaseous and particulate pollutants.
containing respirable droplets that can be inhaled.
arger droplets that are trapped in the upper airway trachea and
°ranchia may possibly diffuse contaminants into the airway.
^Position velocities of droplets are much higher than the gases
Of Particulates themselves, thus enhancing deposition of
w?f .tants onto leaves, plants, crops, and water bodies.
Totting of surfaces such as leaves, plants, forest canopy
increases the ability of intake of pollutants.
223
-------
. fog/cloud could transport toxics to high altitude and
them onto stressed plants/soil* Bus is a concern which I
been evaluated,
. fog occurrence and stagnation can cause local minor and
nontraditional sources of toxics to become a major health
concern.
6. RESEARCH NEEDS
Primarily, receptors that are sensitive to the exposure of
should be identified. The risk for these receptors (human popu
forest, etc.) should be evaluated. Droplet size distribution of
should also be studied to estimate the respirable portion of the
Based on the population, the concentration, and frequency of
health risk can be determined for such sensitive areas. This will
the identification of affecting sources and help reduce the
minimize the risk. Exposure studies with contaminated droplets w
to understand the health effects due to inhalation.
Immediate research needs are: 1) chemical characterization
fog/cloud sanples for trace conpounds in high altitude/pristine
assess its impact, if any, on sensitive forest environments, 2)
characterization of fog samples in a valley which has major source^
toxics .
For exanple, fog in a populated mountain valley under
stagnation conditions, where wood burning is a predominant source?
contain several toxic compounds. A chemical characterization stud/;.,
including an Ames test on the fog droplets, could help to evaluate lW
potential health concerns.
REFERENCES
1) K. Kawamura, S. Steinberg and I.R. Kaplan. Intern, j. Environ
Analytical Chemistry, Vol. 19, 175-188, 1985.
2) K. Kawamura and I.R. Kaplan. Analytical Chemistry, Vol. 56 / &'
Aug. 1984. _
3} D. Grosjean and B. Wright. Atmospheric Environment, Vol. 17' **!
pp. 2093-2096, 1983. ,
4) J.W. Hunger, C. Tiller, M.R. Hoffman. Science, Vol. 231, PP-
249, Jan. 1986. -j
5) J.W. Hunger, D.J. Jacob and M.R. Hoffman. Journal of Atmospkef
Chemistry 1, pp. 335-350, 1984.
6) D.E. Glotfelty, J.N. Seiber and L.A. Liljedahl. "Nature, Vol-
Feb. 1987.
7) D.J. Jacob, J.M. Waldman, J.W. Munger and M.R. Hoffman.
272-285, 1984,
8) J.M. Waldman, J.W. Munger, D.J. Jacob, R.C. Flagan, J.R.
M.R. Hoffman. Science, Vol. 218, 12 Nov. 1982.
9) J.W. Munger, J. William, D.J. Jacob, J.M. Waldman and M.R.
Journal of Geophysical Research, Vol. 88, No. C9, June
10) R.P, Richards, J.W. Kramer, D.B. Baker and K.A. Krieger. N
Vol* 327, No. 6118, pp. 129-131, 14 May 1987.
11) J. Firket. Trans. Faraday Soc., 32, 1192-1197, 1936. »
12) M.R. Hoffman. Environ. Sci. Technol., Vol. 18/ No. 1, 19B4. jf
13) J. Fuhrer. Agriculture, Ecosystem and Environment 17, 153-16*
14) J.M. Waldman and M.R. Hoffman. Advances in Chemistry SerieSr
Washington, DC, 1987.
15) S. Fuzzi and G. Orsi. Journal of Atmospheric Chemistry 3,
224
-------
l6) W n
17) T Oui?rdwick* M0111^ Westher Review 101, 763-766, 1973.
' ^ns- Masters Thesis in Meteorology, United States Naval Post
18) F . Graduate School, Montrey, CA.
19) c*p* Ludgwig and E. Robinson. Tellus 27, 94-105, 1970.
• Junge. Final report, U.S. Weather Bureau Contract No. CWB-
20) E j i1151' Stanford Research Institute, Menlo Park, CA.
•u« Mack, and R.J. Pile. Report No. CJ-5056-M-2, Calspan
21) c B Corporation, Buffalo, NY, 1973.
Kodgers, E.J. Mack and R.J, Pile. Report Mb. CJ-5076-MI,
22) c w Calspan Corporation, Buffalo, NY, 1972.
• Rodgers, E.J. Mack and V. Katz. Report No. CJ-5076-M-3, Calspan
23) D B Corporation, Buffalo, NY, 1974.
• oickson, R.B. Loveland and W.H. Hatch. Capistrano test site
trom 16 April through 11 May. Report No. Ecom-dr 75-3,
»Vospheric Science Laboratory, White Sands Missile Range, MM,
24) D vol. 1, 1974.
' ' of0013 and M.R. Hoffman, Journal of Geophysical Research, Vol.
• Ot> — ------ _w~ ^..mua V WM«. * «*•* ••- Nrf*. Nrf'wVJ.Al.JfM^ %***.!• kVwtd^Ul. WJbl f YWi
5) K.C w N°* CU' PP- 6611-6621, Aug. 20, 1983,
• Weathers, G.E. Likens, F.H. Bormann, J.S. Eaton, W.B. Bowden,
J.L. Andersen, D.A. Cass, J.N. Galloway, W.C. Keene, K.D.
6) W. w.lraba:L1' P. Huth and D. Smiley. Nature, Vol. 319, 20 Feb. 1986.
2 ' loT61' and P* Warneck- Atmospheric Environment 14, 809-818,
28) w! Sj^ean' Envir. Sci. Technol. 16, 254-262, 1982.
. ' icP61 and P> t*1018014' Atmospheric Environment 14, 809-818,
30 ' " 53Qdel and C*J* Weschler' Rev- Geophys. Space Phys. 19, 505-
, ' ' Chaneides and D-.D- Davis. J. Geophys. Res. C87, 4863-4877,
•M r •^•982.
^i *^ W tt •
3i J-A* 5IChards ^-' 3i" Atraos- Environ. 17, 911, 1983.
^ J,N' Jr^dle and M.R. Hoffman. J. Phys. Chem. 87, 5425, 1983.
' r^1:Lowayr G.E. Likens, E.S. Edgerton. Science 194, 722-724,
*) D.j i976«
3 " ' £acob' J«W. Munger, J.M. Waldham and M.R. Hoffman. Journal of
V.E r^^1^1031 Research, Vol. No. Dl, pp. 1073-1088, Jan. 20, 1986.
Thomson, A. Jones, E. Haemisegger, and B. Steigerwald. JAPCA,
W.j JJ1' 33, No. 7, 1983.
E if' ^ai, JAPCA vol. 33 NO. 7 1983.
M!D' S;lber9eld.' JAPCA, Vol. 136, No. 9, 1986.
3\ ' i^ettf W>A' Refers and R.K. Olsen. Science 218, 1303-1304,
' G "°2.
' •* 2°^lard' M»H- Unsworth and M.J. Havre. Nature, Vol. 302, 17
'i 1983.
ACKNOWLEDGEMENTS
was supported by a Fellowship from AAAS/U-S. EPA. I am
,_i "y special thanks to Lisbeth A. Levy of AAAS,
n® Cto-n^tor of this fellowship program. I also would like to thank
lcich, secretary at Montana Tech, for typing this report.
Vftr" grateful to Dr. Glotfelty, USDA Research Station at Belts-
for providing several reference materials and discussion.
225
-------
Table :: Historical Fog Episodes and Associated Excess Deaths
(Hoffman, 1984) ~~
Period of Episode , Excess Deaths
a) London
9-11 Dec. 1873 650
26-29 Jan. 1880 1175
28-30 Dec. 1892 779
26 Nov.-Dec. 1, 1948 800
5-9 Dec. 1952 4000
3-6 Jan. 1956 1000
2-5 Dec. 1957 250
5-10 Dec. 1962 700
7-22 Jan. 1963 700
b) Meuse Valley, Belgium
1-5 December 1930 63 (4-5 Dec.)
c) Donara, PA
27-31 October 1948 20 42.7% fell sid
Table II. Pesticides and Their Alteration Products in Fog Water <6)
Cone, inyg/1
CA Beltsville
Mean* Max
Diazinon 16.00 22.0 ' 0.14
Parathion 23.20 51.4
Chlorphyrifos 2.61 6.5
Methidathion 5.54 15 5
Malathion 0.18 [35 2.74
Methyl parathion - _ 1<21
Parathion Oxygen Analogue 65.65 184.00
Chlorpyrifos Oxygen Analogue 0.49 0.80
Methidalthion Oxygen Analogue 4.16 8.20
Diazinon Oxygen Analogue 0.19 0*19
DEF 0.53 o!sO
Atrazine 0.43 0<70 0<82
Sinezine 0.57 1.2 0.04
Pendunethalin 2.50 3.62
Alachlor - 1 ^ «
Metolachlor - _ 1'g6
Tributyl-phosphate - _ ^g
- 122.95 ug/1
*Mean of three sates; Parlier, Corcoran and Lodi, CA. For detailed
table refer Glotfelty, et al.<6)
Table III. Annual deposition of ions by cloud and bulk precipitation at
Mount Moosilauke, New Hampshire*48).
kg/ha/yr
DC
ion DC* Db** Db
SQ2- 275.8 64.8 4.26
NO'1 101.5 23.4 4.34
^ . 2.4 1.5 1.60 0.62
16.3 4.2 3.88 0.80
5.8 1.7 3.41 0.77
3.3 2.1 1.57 0.61
*Dc and **Db deposition rate due to cloud and bulk precipitation
respectively.
226
-------
IN irr ORGANIC CHEMICAL CHARACTERIZATION OF CLOUDS
HIGH ELEVATION SPRUCE-FIR FORESTS AT MT. MITCHELL, NORTH CAROLINA
Viney P. Aneja and Ronald L. Bradow
Department of Marine, Earth, and Atmospheric Sciences
North Carolina State University, Raleigh, NC 27695
P R. K. M. Jayanty
Kesearch Triangle Institute, Research Triangle Park, NC 27709
1988 EPA and APCA Symposium
°n Measurement of Toxic and Related Air Pollutants
Raleigh, NC
^ tol
5OUt»taiJL0t,clouds and volatile organic pollutants in the high elevation
» tain p? North Carolina is being studied under the auspices of the
k Sttl»»* t d Chemistry Project. Cloud water samples were collected using
» *itch i?% collector during spring, summer, and fall of 1987, at
Jj4ly2ed J11 State Park . ehloroform>. Aromatics ttoluene. o-n-D XVle
n trineJkeihylene cnloride» chloroform), aromatics (toluene, o-m-p xylenes
r* Ptesen De«zenes) at low concentrations (1-B ng/ml). The sources for
*UrtK of tnese compounds in cloud water is speculative at this time
net votk is in progress.
227
-------
INTRODUCTION
High elevation forests (primarily red spruce and fir) in the eastern US *
exhibiting visible symptoms of damage, decreased radial growth, and
increased mortality (Johnson and Siccoma, 1983). Hence there is a l«r*e jj
research effort currently in place to determine causes of forest decline
the US, Canada, and western Europe (Aneja et al., 1988, Saxena et__al»» .1
1988). Under the auspices of the Mountain Cloud Chemistry Project TMcC
it is theorized that cloud deposition at high elevation forests may b*
causing this phenomenon (Valdsterben). However, no cause and effect .
relationship has been established thus far except for the role of oxid»"
ozone.
It is now recognized that no single factor is responsible for forest &£
in North America and western Europe, but a variety of biotic and abiot*c
stresses, including stresses associated with exposure to air pollutant5'^
interact synergistically to bring about this forest decline. Air P°rrrjJ'
act in a variety of ways to alter the physiological processes that un
-------
nr>" known that the atmosphere provides a major pathway for the
. and transformation of pollutants (Duce et al., 1975} Prospero and
*9'7). vhile most of these studies deal with inorganic gaseous and
Pollutants, reports on atmospheric transfer of organics, both
~enic and natural, are sparse. It has been documented that organic
s» namely chlorinated hydrocarbons such as polychlorinated
s Qcic BBO precipn»uM» •»
N Mean
14 044 045 lot.02 16 <0.6
11 0.10 0.095to0.13 W <0.03
9 0.25 0.075100.57 16 3.11 1.31 to 6.13
17 0.015 0.006lo0.021 16 041 0.34 to 149
14 0.013 0.006loO.OI5 16 <0.02
_„,_,- 17 0.010 0.006100.018 16 <0.02
XL"rDE 17 0.003 0.002100.005 16 *0.02
17 OJ7 0.40 tol.W 16 11 a,6«o724
17 M 042 I»24S 16 55 54 to213
(Sourct: Atlas and Giaa (1981).)
***** 3. Comparison of coooentntiont sf tended ot(mnks in the atmotpnere (in nanograms
ger cubic aaeten.
RM. UAHh O"* C0116** Pi«MI1
Cw-jwund .J"*: Jj"* of Station. Key.
**«*t Atlantic Mexfco TeM> ptoria,
^(Aroclorl242) 044 IM <1
0.06 0.69 0.40 0.41
0.003 0.006 0.083 0.34 0.04
0.9 1.0 14 3.8 184
1.4 2.9 1.2 2.4 16.6
0.10 0.15 0.20 0.12
0.25 0.39 > 0.42
0.010 (0.02)* 0.07
0.012 (0.03)* 1.26
(Sourc
e: Atlas and Giam (1981).)
229
-------
Some limited attempts have been made to characterize the organic polly
content in air and rain samples in the remote regions of the world
et al.» 1972} Atlas and Giaa, 19B1). Recently, Glotfelty et al.
reported that a variety of pesticides and their toxic alteration
may be present in radiation fog- However, cloud organic chemical
characterization still remains unresolved, even though clouds provide*
possible pathvay of enriching of atmospheric VOCs and they are transp**
over large distances. Moreover it is now becoming clearer that clouds
the primary deposition pathway of pollutants on high elevation forest
ecosystems.
EXPERIMENTAL
Sampling location/sampling procedures:
The cloud water was collected at Gibbs Peak (-2000 m MSL), located
Ht. Mitchell State Park (Figure 1). This region at high elevation
stands of red spruce and Frazier fir trees (-6 m tall). A 16.5 m
walk-up meteorological tower was installed at the location. It was
at the top with a passive ASRC cloud water collector (Falconer and ,
1980). In addition, Meteorological instruments were placed on the to*1.
"16 m above ground. These instruments included capability for »e«a«- j
temperature, pressure, wind speed and direction, relative humidity, *"
solar radiation.
One hour integrated cloud water samples were collected by the ASRC <
collector. These samples were analyzed for pS within 15 minutes of
collection. Additional samples were stored in sealed 60 ml plastic
In a refrigerator for subsequent chemical cation and anion analysis
utilizing ion chromatography; and VOCs utilizing GC-MS.
Ambient gas phase pollutant concentrations for ozone (0.), sulfur-di01L,
(SO,), and nitrogen oxides (NO, NO , and NO,) were measured continuous
thVsite. * *
Analysis Procedures
Gas chromatography/mass spectroraetry (GC/MS)J The cloud water
initially qualitatively analyzed for VOC content using GC/MS. The - .
analysis of VOCs was performed as follows: an inert gas (He) was bu&
through a 5 ml cloud water sample contained in a specifically design0
purging chamber at anbient temperature. The VOCs are efficiently
transferred from the aqueous phase to the vapor phase. The vapor i*
through a sorbent column (Tenax) where the VOCs are trapped. After P"
is completed the sorbent column is heated (190 C) and backflushed
inert gas (He) to desorb the VOCs onto a gas chromatographic column
meters long, 0.3 mm o.d. coated with DB5. The flow rate through the
is 3 ml/min of He). The gas chromatograph is temperature programmed
for three minutes program to 31°C at 3 C/minute, change the program *v
8 C/min. to a maximum of 200°C) to separate the VOCs, which are then ^
detected with a mass spectrometer. Reagent water blanks were an"1"**
VOCs and found to be insignificant. An aliquot of the sample is
with reagent water where dilution is necessary. A 5 ml aliquot of
dilution is taken for purging.
230
-------
of Mt. Mitchell (MM), North Carolina (NC) research site.
of other stations in the eastern United States operated
under Mountain Cloud Chemistry/Forest Exposure Study Project
IMCCP). VT-Whitetop Mountain, SA-Shenandoah, Virgina (VA)j VF-
^uteface Mountain, New York (NY)j MME-Mt. Moosilauke, New
u«npshire (NH); HL-Howland, Maine (HE).
231
-------
Gas chromatography with flame ionization/electron capture detectors: The
cloud water samples were analyzed for six selected volatile organic
compounds nethylene chloride, chloroform, teluene, ethylbenzene, o-xylene,
and methyl benzene) using a purge and trap concentration and gas
chromatography (GC) with flame ionization and electron capture detection
(FID/ECD). The six compounds vere selected on the basis of GC/MS
confirmatory analysis. Table 3 lists the experimental conditions used for
the analysis of cloud vater samples. Calibration was performed using two
standards containing each of the six compounds prepared in vater at
concentrations of 1 ng/ml and 10 ng/ml. Response factors for the tvo
standards vere averaged and used to calculate the sample concentrations.
Distilled water used to prepare the standards was analyzed each day before
analyzing the samples to check for contamination and found nondetectable
mounts of these components.
Table 3. Analytical Conditions for Cloud Water Analysis
Parameter
Setting
Concentration
Volume of sample
Purge flov rate
Purge tine
Dry purge
Desorb temperature
Desorb time
Focusing *
Gas chromatograph
Column
Carrier gas
Temperature program
Initial temperature
Initial time
Rate
Final temperature
Detectors
Data system
Tekmar LSC-2
5 mL
60 mL/min
10 Bin
2 min
190 °C
4 min
Cryotrapped in liquid nitrogen
Perkin Elmer 3390
60-m, 0.35 mm ID, 1 mm DB-1
coated fused silica capillary
He, at 30 cm/sec
Subambient (N«)
2 min
8 °C/min
150 °C
FID and BCD split 5 to 1
Two channel Shimadzu CR5A integrator
RESULTS AND DISCUSSION
Tables 4 and 5 report results of cloudwater sample analysis for a fev days
in May through October 1987, Chloroform, toluene and the C8 aromatics vere
detected at low concentrations in most samples, while methylene chloride and
mesitylene vere reported as trace levels in a fev, due to the insensitivity
of the detector. The concentrations found for each compound by both GC and
GC/HS are, in general, consistent with each other and wherever the
232
-------
Table 4. Analytical Results of Cloud Vater Samples from Mount Mitchell using GC/FID/ECD.
to
CO
c*>
Methylene
Date
collected
5/18/87
5/21/87
5/22/87
8/13/87
8/19/87
8/19/87
10/6/87
10/12/87
10/12/87
chloride
FID
TR
ND
ND
ND
TR
TR
ND
ND
ND
BCD
ND
ND
ND
ND
ND
TR
ND
ND
ND
Concentration
Chlorofom
FID
7.8
2.0
A. 9
ND
8,3
10
ND
ND
ND
BCD
5.8
2.9
4.2
ND
6.2
€.7
0.35
0.34
0.28
(ng/«L)
Toluene
FID
0.62
0.88
1.0
ND
0.91
0.93
0.36
1.0
1.1
Ethyl-
benzene
FID
0.33
0.23
0.31
ND
0.36
0.45
0.24
0.32
0.29
o-Xylene
FID
0.59
0.39
0.46
ND
0.86
0.78
0.31
0.65
0.49
Hesitylene
FID
ND
0.23
ND
ND
ND
ND
ND
ND
ND
TR * trace levels detected but could not be quantltatea
ND * could not be detected
-------
Table 5. Analytical Results of Cloud Vater Samples fron Mt. Mitchell Using GC/MS.
ISJ
CO
Date
collected
8/06/87
8/19/87
10/12/87
Hethylene
chloride
ND
13
CT
Concentration (ng/mL)
Ethyl- Trims thy 1-
Chloroforn Toluene benzene Xylenes benzenes
ND 1.5
ND 1.9
CT CT
ND 0.85 1.0
0.32 2.6 1.7
CT CT CT
TR * trace levels detected but could not be quantitated
ND * could not be detected
CT =« the sample was supected of being contaminated
-------
1uantif^ °ns are hlBner for few compounds In GC/MS analysis was due to the
er* °n °f a11 the three isomers. Also, there were no significant
er es ln concentrations between the samples collected in three
rent seasons.
Th
j amounts °* aromatic hydrocarbons measured in the gas phase is
* Distorted from urban distribution patterns by chemical reaction
Af sport (Roberts et al., 1984; Singh et al., 1985). Generally, the
th toluene/ethylbenzene~Tn urban air is Between 3 and 5 (Singh, 1985),
ioSatl° of o-xyieoe/etny1060260® ls slightly greater than 1 (Singh
arh At truly remote locations (viz. Nlwot Ridge, Colorado), C8
Cotls«aue iS are considerably depleted relative to toluene or benzene.
*tcas\Bw y reP°rted toluene/ethylbenzene ratios are 10 or greater in such
C°aPound k S g* al" 1984>- In fact» the relative ratios of these
as °ave been used to estimate OH radical concentrations and transport
T^ c
*tu<1y mavnkrations of aromatic hydrocarbons found in cloudwater in this
Iat*t and/ u bccn altered fay the relative solubilities of these species in
! st*ibuM°r y vaP°r pressure considerations, however, the relative
*r
.t ""lor* °?S °f C8s ls very simllar to tnat observed in urban air.
* . **nt £ ls likely that the source of these compounds is not very
Uli* tn* mountain collection site. In fact, the small city of
about it' is a°out 30 k* southwest of the site, and Black Mountain Is
irouoJ: | ** south. It is conceivable that the presence of
d ixnin:c sources from these towns or from US Interstate Highway 40
^ PJ-ain the occurrence of the aromatic hydrocarbons in cloudwater.
J°und i?SJ Preliminary results it is clear that organic compounds can be
. °Porn«°etfctable amounts at mountain summits and that the relative
°* the substances measured may be consistent with relatively
c«s (albeit at very much lower elevation). In future studies it
to'jjturi*0 measure both gas-phase and water-phase substances in parallel
8% form.ij tfte water-phase as a possible concentrator of polar organics,
r"«ldehyde (Adewyi et al., 1984).
•8 t>
CK^^nmenf0!) nas been funded through a cooperative agreement with the US
*P?isttv J Protection Agency (813934-01-2) as part of the Mountain Cloud
»Kr°vUdi» 2gram» Dr* Volker A. Mohnen, Principal Investigator. We
aytUal Caro1 Haney» Dr- A* R« Gholson, And Mr. S. B. Balik for
ia.1 -SuPport. Continual technical discussions with Dr. R. Paur,
Saxena, and Dr. Ellis Cowling are greatly appreciated.
ksf ?f this document do not necessarily reflect the views and
Nts of »k Environmental Protection Agency, nor the views of all
<«t s Or c Mountain Cloud Chemistry consortia, nor does mention of trade
or non-commercial products constitute endorsement or
for use.
235
-------
REFERENCES
Aneja, V. P., C. S. Claiborn, S. R. Chiswell, R. L. Bradov, R, J. Paur
R. E. Baumgardner. 1988. "Dynamic Chemical Characterization of
Clouds", Air Pollution Control Association. 81st Annual Meeting,
Texas *
Atlas, E. and C. S, Claw. 1981. "Global Transport of Organic
Ambient Concentrations in the Remote Marine Atmosphere11, Science, Vol-
pp. 163-165.
Bevenue, A., J. N. Ogata, J. V. Hylin. 1972. Bull. Environ. Contain.
Toxicol., Vol. fl, pp. 238.
Bidleman, T. P., C. P. Rice and C. E. Olney. 1976. Marine Pollutant
Transfer, fl. L. Windom and R. A. Duce, eds. (Lexington Books, Lexi
Mass.), pp. 323-351.
Duce, R. A., G. L. Hoffaan and V. H. Zoller. 1975. Science, Vol.
pp. 59.
Falconer, R. E. and P. D. Falconer. 1980. "Determination of cloud
acidity at a aountain observatory in the Adirondack Mountains of NeV tg
State", Journal of Geophysical Research, Vol. 85, No. C12, pp. 7465*7*'
Giaa, C. S., M. S. Chan, G. S. Neff and E. L. Atlas. 1978. Science* *° '
199, pp. 419.
Glotfelty, D. E., J. N. Seiber and L. A. Liljedahl. 19B7. "EeaticH**
fog11, Nature, Vol. 235, pp. 602-605.
Johnson, A. B. and T. A. Siccama. 1983. "Acid deposition and fores*
decline", Env. Sci. & Tech., Vol. 17, pp. 294-306.
Mohnen, V. A., J. Bealey and B. Bailey. 19BB. "Exposure of forests *°
gaseous air pollutants, clouds, and climatic variables", Atnospheric
Chemistry Task Group II, NAPAP, Research Summaries.
Murphy, T. J. and C. P. Rzeszutko. 1977. J. Great Lakes Res., Vol' 3f
pp. 305.
Prospero, J. K. and R. T. Nees. 1977. Science, Vol. 196, pp. 1196*
Roberts, J. H., R. S. Hutte, F. C. Fehsenfeld, D. L. Albritton and
R. E. Sievers, 1985. "Measurements of Anthropogenic Hydrocarbon R
the Rural Troposphere", Atmos. Environ., 19, 1945.
•p*
Roberts, J. M., F. C. Fehsenfeld, S. C. Liu, M. J. Bellinger, C-
Albritton and R. E. Sievers. 1984. "Measurement of Aromatic
Ratios in the Rural Troposphere", Atmos. Environ., 18, 2421.
Saxena, V. K., R. E., Stognert A. B, flendler, T. P. Defelice, J. ?« t
H. Lin. 1988. "Monitoring the chemical climate of the Ht. Mitch*!1 L
Park for evaluating its impact on forest decline", Tellus 41B, in Pre
' " h J* Salasp B' Kl Cant«H and R. «• Redmond. 1985
°f Aromatic Hydrocarbons in the Ambient Air", Atnos.
236
-------
ft
Ca atlve Importance of Dry, Wet, and Cloud
PtUre Mechanisms for Acidic Deposition
pa aandN--H.Lin
n ment °f Marine> Earth and Atmospheric Sciences
Carolina State University
lgh'NC 27695-8208
Th
tp°riQech ^nt* ^ deposition °f acidic substances at the forest canopy are considered as
.S for P°llutant induced forest decline in high elevation mountains in eastern
above cl •• Direct cloud capture plays a predominant role of interceping acidic substances
- ' 1986* se forests- Cloud-intensive observations were initiated at Mt. Mitchell in
tic1>n orc*er to quantitatively assess the exposure of forested ecosystems to
jj lnPuts and to determine the relative importance of wet, dry and cloud capture
lpi' According to the database obtained in 1986 and 1987, Mt. Mitchell was exposed
be th ^^ °f days per year. Sulfate, nitrate, ammonia and hydrogen ions were
i n*v)C P"nc*Pal *ons *n the cloud water. Using a micrometeorological cloud
ntrih ^ ^e annua* cloud water deposition was estimated to be 14.7-26.7 cm yr"1.
.^ 3,:u?d to ^ total deposition about 2-3 times more than the nitrate. Based on the
,j of these two years of database and the presumed rate of cloud water deposition
,r* I ^ indicator of sulfate deposition vs. other major ions is suggestive of 1 and 2
, the m>SU!fate corresponding to the cloud water pH=3.67 and 2.88, respectively,
?e events with precipitation were found only with average pH values around
general. \ve> therefore, suggest that the direct cloud capture mechanism is
roost acidic deposition in high elevation ecosystems such as the.ones existing at
237
-------
1. Introduction
During the last three decades, the chemistry of precipitation has been widely investigated in
many industrialized areas, such as western Europe, northeastern United States and Canada,
which are significantly affected by the acidic precipitation that has enhanced acidification of
lakes and plant-soil systems, as well as linked ecosystems. Junge1 concluded that the
hydrogen ion concentration in rain is not a diluted aerosol component, but is the result of
fixation of SO2, NO2 and HC1. It was further argued2 that precipitation obtains a substantial
portion of its acid by incorporating gaseous sulfur and nitrogen compounds via transformation!
involving liquid water in the atmosphere. From these arguments it can be inferred that clouds
play a relatively important role in the transformation and deposition of acidic substances. The
previous investigations on cloud and fog water chemistry supported this point. However, the
attention paid to the cloud chemistry was relatively limited, but the studies in chemistry of
airborne clouds and cloud water collected using ground based collectors are now available.
Although the pH values and chemical composition of cloud or fog water varied in space and
time, the former were substantially lower than the CO2-equilibrated value of pH=5.6 ovei
industrial regions and even over remote areas free from pollution in their immediate vicinity,
and the latter were principally found to be dominated by sulfate, nitrate, ammonia and
hydrogen ions.
Saxena et al. 3'4 have pointed out the relative importance of dry and wet deposition as i
possible cause of forest decline in forested mountain areas, which recieve considerable occult
precipitation through horizontal cloud interception. It was further indicated that direct
interception of cloud droplets is the major mechanism of acidic deposition in cloud-capped
mountain areas. For Instance, M» Mitchell (2,038 m MSL, 35°44'05"N, 82017'15"W),
North Carolina,is exposed to cloud and fog episodes during 71% days per year, on the
average. In the above area, the decline of red spruce and Fraser fir forest is noticeable above
the cloud base which is frequently observed around 1,585 m MSL. In this paper, we repon
the results of our studies of cloud water chemistry and using a micrometeorological model,
assess quantitatively the acidic deposition on mountain top due to cloud capture mechanism,
based on the database for Mt. Mitchell for the summers of 1986 and 1987. Saxena and
Stogner5 have used a limited number of 1986 cloud episodes in presenting their estimation of
deposition fluxes of cations and major acidic anions. In the following, a broader database is
used including 49 episodes in 1986 and 44 in 1987.
2. Experimental
A 16.5 m tall walk-up tower was fully instrumented with auto sensors simultaneously
measuring the meteorological parameters which were sequentially recorded on 15 minutes
average of 5 second samples with a tape-trasferred data logger. The manual cloud water
collection commenced hourly at the tower top upon a cloud event inception, using a
teflon-string passive cloud water collector designed and fabricated by the Atmospheric Sciences
Research Center (ASRC), State University of New York at Albany. The pH and total
collected volume of the sample were immediately measured within 15 minutes after collection,
The ion composition of cloud water was analyzed using the conventional ion exchangt
chromatography within one week after sampling. The cloud water samples were kept
refrigerated in the intervening time.
A summary of the observed cloud events and the collected samples is given in Table 1,
The corresponding accumulative and average duration of events are listed in Table 2. Cloud
events are categorized into two classes, long cloud events with a duration exceeding 8 h and
short cloud events lasting no more than 8 h. In general, the former occurred as a result of the
frontal passages and the latter due to orographic lifting mechanisms. In addition, cloud events
with precipitation are categorized as mixed events.
The ion deposition flux is computed from the product of mean ion concentration and the
hydrologic flux in terms of the rate of cloud water deposition. At present, there is lack of
238
-------
Table 1 Summary of events and samples collected at Mt. Table 2 Accumulated duration and average duration of cloud
Mitchell site during the summer of 1986 and 19 87. events.
Event Type
tong
Start
Mked
(nth precipitation)
Total
1986
events / samples
13
26
10
49
94
55
53
202
1987
events / samples
5
32
7
44
46
40
40
126
Event Type
Long
Short
^ed
(wilh precipitation)
Total
1986
accu. / ave.
170.50
92.92
85.82
349.24
13.12
3.57
9.54
7.23
1987
accu. / ave.
53.00
80.40
47.83
181.23
10.60
2.52
6.83
4.12
effective instruments to directly measure the rate of cloud water deposition. Recently, a
disposition model was proposed by Lovett6 and was further modified by Mueller7. This
model consists of two submodels: 1) structure submodel simulating the structure of forest and
the growth of trees, and 2) hydrology submodel computing the evaporation rate and the rate of
cloud water deposition resulting from the cloud droplet impaction and sedimentation. Here,
we implement this model to compute the rate of cloud water deposition with the two year
database of Mt. Mitchell, and thereby, extrapolate the annual acidic deposition contributed by
direct cloud capture mechanisms.
3, Results and Discussions
a Mountain Cloud Water Chemistry
Table 3 gives the summary of cloud water chemistry with respect to category of event types.
The calculated pH values were found to be correlated with field measured pH on the 97% level.
The balance between anions and cations was within 5% error. Noticeably, the pH in cloud
events were much lower than in mixed events. It is suggested that acidic substances are
Muted by precipitation in mixed events. Precipitation below the pH range of 4.0 to 4.7 is
regarded as damaging to sensitive aquatic ecosystems8. Some exposure studies have reported
ihe cases of specific injury and growth retardation for several plant and tree species linking the
effects generally noted9'10 in the range of pH=2 to 3. By contrast, our observations show
that the prevailing cloud water pH value is aroud 3.5 in Mt. Mitchell forest area, on the
average. The principal ions are sulfate, nitrate, ammonia and hydrogen. Figure 1 illustrates
die results of linear regression analysis between sulfate and the other ions or their combination,
and the corresponding regression equations. According to our database, sulfate is much better
correlated with cations. As expected, sulfate together with nitrate can account for above 95%
of cloud water acidity. Besides, nitrate is found on the same level with ammonia. Marine
substances contributed less than 5% of the total ion concentration.
i. Acidic Deposition Fluy
The results of model runs are listed in Table 4. Because of insufficient measurements of
liquid water content, a range of 0.01 to 0.4 g m"3 was presumed for 1986 database in order to
make compatible runs of the model. This presumed range with a median of 0.205 g m"-* is
somewhat consistent with that of 1987 data. Further, the annual rate of cloud water deposition
is extrapolated on the basis of the mean cloud duration listed in Table 2 and under the
assumption that 71% of cloud frequency is realized in our case,. As a result, the annual cloud
durations provide approximately 20% (1750 h) and 10% (870 h) immersion time per year, for
1986 and 1987, respectively, and are well compared with 15% for the Smoky Mountain site in
Tennessee.11
The mean and annual ion deposition fluxes, summarized in Table 5(a) and (b), respectively,
for 1986 and 1987, are computed with above parameters and the data in Table 3. Evidently,
sulfate and nitrate were the major contributors to the total mass flux of ions. Gorham et al. 12
have demonstrated that the mean sulfate deposition flux due to wet deposition with annual bulk
precipitation 87.7 cm yr'1 in estern United States, is 23.1 kg ha"1 yr1 corresponding to the
raeanpH value of 4.3. In contrast, our observations show that with annual rate of cloud
239
-------
_
T**"*18* of hourly mean cloud deposition nut and
""Potated yearly rala for ML Mitchell, 1986 tod 1967,
Evf|lt TYM
Long Short
(mm IT1)
0.005-0.304 0.005-0.302
0.03 - 0.4 0.03 - 0.4
0.115
0.217
Yearly pirtimy^
Long Short Total
(cmyr'1)
0.6-34.0 0.3-18.4 0.93- 52.4
8.3
6.4
14.7
*FSSplJ!!"*'ure<)u»'»gFSSPaiid were madeby Saxeu and Stooger*. In 1987.
f!|*__ — ™ "v»w ••••Mfc UWb IV MLnUMU UIUUKUEI UU
"^method provided by T«nneMoo
Sulfatt Depotltion (cq, hr1 «•')
246
• I*IM*VULW In UUI
.IJ>ltOr>],|«Jf IT. MM)
• U>|tOr>|.HUI In M» ,/ , L2
L1
»00 2000 MOO
Sulfkte CooentnUoa (Mcq. f')
4000
water deposition of 14.7 cm yr"1 only
and the mean pH value at 3.5 level.
the sulfate deposition was 26 kg ha'*
yr1 for 1987. Evidently, direct cloud
capture mechanisms contribute to
acidic deposition substantially,
especially in places like Mt. Mitchell
State Park.
We have demonstrated in Fig. 1 the
estimation of ion concentration due to
sulfates. Here, we further assume a
prevailing rate of cloud water
deposition, 0.2 mm h'1, for running
the cloud deposition model, and
construct an indicator of predicting
ion deposition flux, which is dipicted
along the upper abscissa and right
ordinate in Fig.l. As the pH value
of cloud is measured, the
corresponding sulfate concentration
associated with deposition flux, as
well as that of the other ions or their
combination, can be derived from
Fig. 1. The usefulness of the
representation is self-evident.
°f fceeKinuu, of Ionic nun fliua ta ML Mitchell m*. 19*6.
Loot
YHdz.
Short
-------
Acknowlegdements
The research has been funded through cooperative agreements with the U.S. Environment
Protection Agency (agreements No. 813934-01-2 with the State University of New York
Albany and contracts CRS 812444-01-0, 02-1, and 03-0 with the North Carolina Stale
University). Professor Volker Mohnen is the Principal Investigator on the Mountain Cloiii
Chemistry Project (MGCP) and the EPA project officers are Drs. Ralph Baumgardner aui
Richard Paur. The contents of this paper do not necessarily reflect the views and policies ci
the EPA, nor the views of all members of the MGCP consontia, nor does the mention of trade
names or commercial or non-commercial products constitute endorsement or recommendatio:
for use.
References:
1. C. E. Judge. 'Air chemistry and radioactivity.' Academic, New York, 3SS (1963).
2. L. Newman, ' General consideration of how rainwater must obtain sulfate nitrate
and acid,' Chemical Congress. Am. Chem. Soc.. Chem. Soc. of Japan. (1979).
3. V. K. Saxena, P. Agarwaal, S. Raman, ' Wet and dry deposition of air pollution or
the forest canopy as a cause of forest decline,1 Conference on Tropica'.
Micrometeorologv and Air Pollminn. New Delhi, India, Feb. 15-19 (1988a)
260-265.
4. V. K. Saxena, R. E. Stogner, A. H. Hendler, T. P. DeFelice, J. -Y. Yeh, N. -R
Lin, Monitoring the chemical climate of the Mt. Mitchell State Park for evaluating in
impact on forest decline,1 Tellus. 41B (1988b, in press).
5. V. K. Saxena, R. E. Stogner, 'Wet deposition on forest canopy at Mt. Mitchell,
North Carolina,' In: Measurements of Toxic and Related Air Pollutants. Pittsburgh,
APCA, 189-194 (1987).
6. G. M. Lovett, ' Rates and mechanisms of cloud water deposition to a subalpine
balsam fir forest,' Atmos. Environ. 18: 361-371 (1984).
7. S. F. Mueller, ' Chemical deposition to high elevation spruce-fir forests in the
eastern United States,' A Prelimilarv Assessment to Mountain Cloud Chemistry
Project (1987), available from the Tennessee Valley Authority.
8. A. Henriksen,' A simple approach for identifying and measuring acidification of
freshwater.' Nature 278: 542 (1979).
9. B. Raines, M. Stefani and F. Hendrix, 'Acid rain: Threshold of leaf damage in eight
plant species from a southern Appalachian forest succession,' Water Air, and So
Pollut. 14: 403-407 (1980).
10. T. Scherbatskoy and R. M. Klein, ' Response spruce and birch foliage to leachinc
by acidic mists,1 J. Environ. Oual. 12: 189-195 (1983).
11. S. E. Lindberg, D. A. Schaefer, J. G. Owens and D. Silsbee, ' Integrated forest
study annual review summary,' Environmental Sciences Division, Oak Ridge
National Laboratory, January 1988,available from Dr. S. E. Lindberg, ORNL.
12. E. Gorham, F. B. Martin and J. T. Litzau, 'Acid rain: Ionic correlations in the
eastern United State, 1980-1981,' Science 225:407-409 (1984).
242
-------
IN CLOUD HATER
B Institute of Technology
React°r Laboratory
Street
Massachusetts 02139
n n""c3-0u^ processes are considered to be one of the major factors
°it th 4 e atmospheric heterogeneous chemistry, very little is known
t^nts r CoiDP0sition except for some major anions and cations. Trace
80ur' ^^es their possible catalytic functions, are fingerprints for
?ater, Cea ^nfltural and antropogenic) of the material found in the cloud
t Cotlda k teen cloud water samples collected at Whiteface Mountain,
j °i» ail . s» New York, have been analyzed by instrumental neutron activa-
u^catarf ^8^8* Sample to sample changes of trace element concentrations
^ft tft that the differences in air masses entrained in clouds are the
8°n for this observed variation.
243
-------
INTRODUCTION
Numerous receptor modeling studies have shown that elemental compost
tion of environmental samples provides invaluable information about th
pollution sources or source regions and their contributions at a receptor
site. These studies are carried out on local, regional and even gloU
scales1'3. In spite of its success, one of the major uncertainties In thU
approach is the lack of information about the in-cloud processes where eojl
of the heterogeneous atmospheric reactions take place. Clouds modify*
original atmospheric composition during the condensation nuclei formation,
growth, and fall as rain. Also, until wet precipitation occurs there..;
be several condensation-evaporation cycles enhancing the chemical react OB
and secondary particle formations. Although these facts are acknowl dg*
in a number of reports and research articles, still very little is kno«
about,the elemental composition of cloud water.
There are several hundred publications on the chemical composition oi
rain and/or snow, but until recently little attention has been given to to
chemistry of cloud water, ground fog, mist or other forms of water In tk
atmosphere. Clouds or fog were analyzed primarily for their major lodt
content, such as H+, Nu£. SO,,, N03, etc., without ™**~*££\m™
elements (see extensive bibliographies given by Jacob and Hoffman, Muir ct
al.5).
In studies of the origin and fates of trace elements in the environ-
ment, it is advantageous to be able to analyze samples for a wide spectra
of elements with a high sensitivity and accuracy. Although this cond tl«
is best satisfied with instrumental neutron activation analysis (INAA),«
found only one other study, that of Bogen,6 who applied this technique t.
analyze cloud water samples. He found some of the elements (Ag, Sb, Co, I
and CD were enriched in these samples compared to the average crust.]
abundance pattern.
EXPERIMENTAL
In this study we analyzed fifteen cloud water samples by INAA tt
determine their elemental composition and their relations to anthropogenk
and natural sources. Samples of cloud water were collected by Kadlecek *
his group at Whiteface Mountain, Adirondacks, New York, during an extensile
cloud period between May 22-24, 1987. In this period a weakening low pres-
sure system moved from Minnesota to northern Michigan. It passed just U
the north of Whiteface Mountain and into the Atlantic. A listing of A
samples which represent cloud water from both the warm and cold sectors i
conjugate frontal passages with starting and ending dates and times, th
presence of precipitation and the results of field PH measurements at,
given in Table I.
In order to Increase the sensitivity and eliminate geometry facton
during counting, about 50 ml of cloud water was freeze-dried individua y.
Five ml aliquotes were put into previously acid-cleaned small polyethylw
bags (-50 mg). Following the initial freeze-drying, additional aliquot.
were added to the same bags and the process was repeated. Then the bag.
containing the residue were heat-sealed and placed into another clean bq
244
-------
to prevent any surface contamination during irradiation. Concentrations of
the elements were determined by irradiating the samples at the MITR-II
nuclear research reactor and measuring the y-rays of radioactive isotopes
produced as the result of neutron irradiation. The INAA procedure applied
was very similar to those reported previously on other environmental sam-
ples.7
RESULTS AND DISCUSSION
Concentrations of the elements determined for each cloud water sample
are given in Table II. Ten elements were observed in at least seven sam-
ples. Some others, such as S, Cr, Fe, I, La, Sin, Eu, W, and Hg, either
vere observed in a few samples or their concentrations were not deter-
ilned. Since this study, we have modified our irradiation and counting
procedures and are able to determine concentrations of more than twenty
elements in cloud and rain water.8 >9
Examination of the elemental concentration patterns and the meteoro-
logical conditions indicated two distinctly different cloud events* The
first one, which was associated with a warm front (samples 1-8), represents
air parcels originating from the Great Lakes region. The second, a cloud
event (samples 9-15) with a shallow layer of cold air, represents air
lasses which were traveling from southern Quebec. There were significant
differences between these two events with respect to loadings of, espe-
cially, pollution-related elements and pH.
Variations in the absolute concentrations of elements observed from
one cloud event to another can be due to various factors. Two of the most
inportant ones are: changes in the liquid water content (LWC), and differ-
ences in the material loadings depending upon the origin of air masses.
Since the LWC was not measured during the sample collection, we are not
able to assess its effect. But if the variation in LWC was the major fac-
tor in observed concentration differences, we would expect that it should
affect all of the elements similarly. Instead, examination of the data
presented in Table II reveals that there are three groups of elements
behaving differently (Figures 1, 2a and 2b). Group No. 1, which includes
elements mainly related to coal combustion, As, Se, Sb and Zn, shows a very
jtrong correlation with [H+] (Figure 1). The other two groups (Figures 2a
aid 2b) have less pronounced correlation with [H+], and high elemental con-
centrations are observed in different cloud water samples than those of
)roup No. 1 elements. Vanadium, which is always associated with oil com-
tostion, has its maximum concentration value at sample 7, along with Mn and
Jr. Calcium, although there are a limited number of observations, does not
show a strong variation and correlation to any other element. Chlorine,
Alch can be related to marine aerosols and incinerators, has a completely
different pattern, with maximum concentrations at samples 8 and 9.
Correlation of these results with our additional studies on precipi-
tating cloud events, coupled with meteorological and ion data,
, SO*, NOj, NH4, etc.)9 indicates that the second factor, i.e. differences
ID air masses entrained in cloud, is the dominating factor on observed
variations in the concentrations of elements. Further systematic studies
ire needed to confirm this finding and to understand more about the
imposition of condensation nuclei.
245
-------
ACKNOWLEDGEMENTS
Contributions and help by J. Kadlecek (WHO), M. Hayes, and MU**1
reactor personnel are gratefully acknowledged.
This work was supported In part by the U.S. Department of Energy
Grant Mo. DE-FG-2-80ERI0770. However, any opinions, findings, con
or recommendations expressed herein are those of the author and
necessarily reflect the view of DOE.
REFERENCES
1. T. G. Dzubay, R. K. Stevens, G. E. Gordon, I. Olmez, A. E.
W. T. Courtney, "A composite receptor method applied to Philadfl
Aerosol." Environ. Sci. Technol.. 22: 46 (1988).
2. S. G. Tuncel, I. Olmez, J. R. Parrington, G. E. Gordon, R' f
Stevens, "Composition of fine particle regional component in S^
doah Valley," Environ. Sci. TechnoU. 19; 529 (1985).
f rfP
3. K. A. Rahn, "The Mn/V ratio as a tracer of large-scale sources of r
lution aerosol for the Artie," Atmos. Environ.. 15: 1457 (1981)*
~~ / />
4. D._J. Jacob, M. R. Hoffman, "Dynamic model for the production <**
N03, SO^ in urban fog," J. Geophys. Res.. _88: 6611 (1983).
5. P. S. Mulr, K, A. Wade, B. H. Carter, T. V. Armentano, R. A.
Fog chemistry at an urban midwestern site," J. Air Pollut^
Assoc.. ^6: 1359 (1986).
6. J. Bogen, " Trace elements in precipitation and cloud water
area of Heidelberg, measured by instrumental neutron activation
sis," Atmos. Environ. 8: 835 (1974).
"~
7. M. S. Genoani, I. Gokmen, A. C. Sigleo, G. S. Kowalczyk, !•
M. Small, D. L. Anderson, M. P, Failey, M. C. Gulovali,
quette, E. A. Lepel, G. E. Gordon, tf. H. Zoller, "Conce
elements In the National Bureau of Standards' bituminous and
minous coal standard reference materials," Anal. Chem.
(1980). "*
8, I. Olmez, "Elemental composition of cloud-rain systems: Clo^d >,•{#
seption, or rain - which has more adverse effects on vegetati°°
preparation).
«&
9. I. Olmez, G. J. Keeler, M. Hayes, K. D. Kimball, J. Kadleceki f
contributions to the chemical composition of cloud and rain I**
preparation).
246
-------
TABLE I: List of cloud water samples analyzed by INAA
Sple
1
2
3
4
5
6
7
8
9
10
11
12
13
U
IS
ntwber and date
- 22MAY87
- 22KAY87
- 23MAY87
- 23MAY87
- 23MAY87
- 23MAY87
- 23MAY87
- 23MAY87
- 24MAY87
- 24MAY87
- 24MAY87
- 24MAY87
- 24MAY87
* 24MAY87
- 24MAY87
Start time
21:00
22:00
04:04
05:04
08:47
09:49
16:49
23:49
00:58
08:54
11:30
12:30
15:30
16:30
17:30
End time
22:00
23:00
05:04
06:04
09:49
10:49
17:49
00:58
01:58
09:54
12:30
13:30
16:30
17:30
18:30
Precipitation
Very little rain
Very little rain
None
None
None
None
None
None
None
None
None
None
None
None
None
_PJL
3.55
3.63
3.87
3.75
3.48
3.35
3.43
3.89
4.00
4.73
4.53
4:43
4.46
4.43
4.59
247
-------
ro
A
to
[
123456789 101112131415
Sample Number
LEGEND
[H] *2E3
As
Se
Sb
Zn/100
FIGURE 1: Concentrations of hydrogen ion and trace elements in sequential samples
of cloud vater.
-------
FIGURE 2a
P
P
b
1234
6789 101112131415
Sample Number
* LEGENCL,
[H]
•Mn
•Br
•V*10
FIGURE 2b
* LEGEND.
23456789 101112131415
note in #1. #2 Na/2
Sample Number
CH]
•ci
•Ca/10
•Na
FIGURE 2: Concentrations of hydrogen ion and some other trace
in sequential samples of cloud water.
250
-------
PLICATION OF CRYOGENIC TRAPPING AND TWO-DIMENSIONAL
o5v*CHROMATOGRAPHY FOR THE MEASUREMENT OF ATMOSPHERIC
4IGENATED HYDROCARBONS
Lin, David Pierotti* and Miriam Lev-On
Environmental, Inc.
, California 93010
^Low molecular weight oxygenated hydrocarbons such as aldehydes,
and other carbonyls are trace atmospheric constituents that play
PoliJOr r°le *n the chain of reactions occurring in both the natural and
°*id €d atmosphere. These compounds are either formed by photochemical
*Uto *on °* hydrocarbons or emitted directly from combustion processes,
r>!!! ,^e exhausts, and biogenic sources. However, the methods
used for sampling volatile air contaminants are not always
to the measurement of oxygenated hydrocarbons at the trace
tired for the understanding of atmospheric processes. The main
"s encountered include poor sensitivities and low recoveries from
3s *8 media or canisters due to decomposition and/or irreversible
°rPtion on surfaces.
fan
&
method was developed to measure low molecular weight oxygenated
Carbons at ambient concentrations. This method uses cryogenic trap
v^th a two-dimensional gas chromatographic technique for separation
?Ua«titation. A detection limit of 10 pptV in A liters of air
a is attainable. Two-dimensional gas chroma tography enables the
Pfobi tlon of tar8et compounds from water, which usually poses a serious
c* ^or t^ie analysis of. polar compounds. This method is also appli-
for the sampling and analysis of light aromatic hydrocarbons.
This technique was used during the summer of 1987 at two of the
n Calif<>rnia Air Quality Study (SCAQS) sites. The target com-
^nvestigated include: Acetaldehyde, Propanal, Butanal, Acetone,
olein, Methylethyl Ketone, Methyl vinyl Ketone, Benzene, Toluene
prene . The identification of these compounds was confirmed by
e analysis of duplicate samples. Data collected from a total of
sampling days during the summer study will be presented.
-
th: Dept. of Earth and Planetary Sciences
Harvard University, Cambridge, MA 02138
251
-------
INTRODUCTION
The importance of low molecular weight carbonyls (C--C,) such as
aldehydes and ke tones, as key intermediates in atmospheric processes
has been discussed extensively before (1,2). The primary sources of
atmospheric carbonyl compounds are:
* Primary emissions, mainly as residues of incomplete
combustion; and
* Intermediate compounds in the photooxidation of organic
atmospheric constituents.
They play a major role in the chemistry of trace atmospheric c°nstis,
tuents by being a source of free radicals due to their ease of photolX9
and by interaction with particles of condensed matter such as soot
adsorption or rain solubility, due to their polarity.
Sampling and analysis of airborne carbonyls are limited due to th«
high solubility in water. Currently, the most commonly used method *°r
the measurement of carbonyls in the ambient air involves the collecti°n
of air samples with a selective reagent such as 2,4-dinitrophenyl-
hydrazine (DNPH), coated on a solid sorbent cartridge or in an impinff®
solution. In both cases, the reaction forms 2,4-dinitrophenylhydrazon«
derivatives which are consequently analyzed by HPLC-UV (3). However*
this method suffers from low sensitivity (0.2 ppbV for a 100 liter a*r
sample), impurities in the reagent and long sampling hours (1-3 hr)«
A well-developed two-dimensional gas chromatographic technique usl
a packed column and a capillary column was employed in this study to
analyze carbonyls in air. This technique coupled with cryogenic- trap
sampling has been used very successfully for the determination of
background ambient levels of carbonyls at industrialized and remote
(4). The advantages of this technique are: short sampling time,
aircraft sampling is possible; separation of water from the carbonyls j,
a packed column prior to analysis; low detection limits which are su*1
for background measurement, and no need for derivatization.
This study was conducted at the Claremont and Long Beach sites t°
currently with the 1987 Southern California Air Quality Study ( SCAQS) •
EXPERIMENTAL METHODS
Sampling Sites, Periods and Days — Twd sampling sites were seJ
for the study. One was located at the SCAQS Claremont trailer and
other was located at the SCAQS Long Beach station. Sampling was cono
at the Claremont site on the 19, 24-25 of June, 13-15 of July, and
of August. At the Long Beach site, sampling was conducted only on « .
of August for comparison purposes. Sampling periods, chosen to coinc* ^
with other sampling activities at the sites, were 5:30, 7:15 and 9i*-V:.
and 12:45, 2:30 and 4:20 PM on each sampling day. Usually, the a
sampling time was less than 5 minutes, and a total of 4 liters of
sample was collected.
252
-------
e Target Compounds — The method used in this study cannot detect
50ttoaldehyde. Formaldehyde is discarded along with light hydrocarbons
,* the pre-column separation. Therefore, the compounds of interest
or, ed: acetaldehyde (ethanal), propanal, butanal, methacrolein (MAC
/jj 2-butenal), acetone, methylethyl ketone (MEK), methylvinyl ketone
W»K), benzene, toluene and isoprene.
Sampling System — The sampling system consisted of four components:
sampling line made of Teflon tubing, a 1/4" O.D. "U" shape Pyrex glass
£FaP filled with 1 mm diameter glass beads and Teflon wool plugs, a Dewar
ias*t containing liquid Argon and a stainless steel diaphragm pump.
j Air Sampling — During sampling the system was purged by ambient air
?r several minutes, then the flow was diverted, by a 3-way valve, while
sampling trap was immersed in the liquid Argon. When the sampling
cooled down to cryogenic temperature, the flow was redirected
th ^ tne traP» and a Measured volume of air was sampled. In general,
SamPllnS flov vas 2~3 liter per minute, and the total volume of air
was 4 liters, as measured by a dry test meter.
Analytical System — The pre-column was a 4 ft x 4 mm I.D. glass
packed with 15* (W/W) BCEF [N,N-Bis(2-cyanoethyl) formamide] on
mesh Gas Chrom Q II with a helium flow rate of 30 ml/min. The
ical column was a 25 m x 0.32 mm I.D. DB1 J&tf Scientific capillary
for n> w*tn a heliura flow rate °f * ml/min• The temperature programs
Analytical sequence were as follows:
Temperature Events
100°C (Trap is thermally desorbed onto
pre-column)
•2 min 30 Dump light hydrocarbons
•9 min 30 Freeze-out the effluent from
pre-column
h min 30 Dump water and heavy hydrocarbons
min - (Thermally desrob cryofocusing materials
from interface onto analytical column)
L9 min 10°C/min Start temperature programming on both
columns
100 Isothermal
Compound Confirmation by GC-MS — Duplicate samples were collected
^alyzed by both GC-FID and GC-MS. A Finnigan OVA GC-MS system
Ped witn tne valve system for two-dimensional gas chromatography
d for compound confirmation. Due to the low molecular weight
tar8et compounds, four mass/charge ranges were selected to avoid
rences from nitrogen, oxygen and carbon dioxide in the back-
The M/Z ranges selected are 19 to 27, 29 to 31, 33 to 43, and 45
ij Tfte standard runs were performed by direct injection of gas
ds Prepared in a static dilution bottle. The identification of the
r* compounds was confirmed by the user's library. Tentative compound
^ification was performed by comparing the mass spectra of the samples
of the NBS library.
253
-------
RESULTS AND DISCUSSION
A total of 54 samples were collected and analyzed from the ,-
site, and 12 from the Long Beach site. In addition, 6 duplicate satnplcS
were collected for GC/MS confirmation.
For all samples, seven carbonyl compounds, two aromatics and
isoprene were identified and quantitated. Formaldehyde could not be
measured due to limitation of the analytical system. The GC/MS
tion of compound identity has included: acetaldehyde, acetone,
butanal, MEK, benzene and toluene. The GC/MS analysis also provided
information on tentatively identified compounds such as methylene
chloride, trichloroethylene, hexane and some other saturated and
unsaturated aliphatic hydrocarbons.
The range of concentrations observed at the Claremont site,
the nine sampling days of field measurements, are presented in Table *•
During the period of September-October of 1980, Grosjean had measured tn
concentrations of carbonyls in ambient air at the Claremont site, usi"*
the DNPH-HPLC method (5). The results of that study are also shown i"
Table' 1 for comparison purposes.
Although acetone was present in many samples in the 1980 study» |*
could not be quantitated due to some background impurity in the samplin|
reagent. Nonetheless, the concentration ranges of the carbonyls measU**
here are quite comparable to those of the 1980 study, although the PrC*^
vious study showed consistently higher concentrations than those °bser!t
in 1987. This may be due to the short sampling time used in the PreSSjffl
technique as compared to the 1 to 3 hours of sampling when using the >&
method. Also, the differences might be attributable to the atypical*?
low seasonal temperatures during the 1987 study.
The average observed concentrations of the measured carbonyls, . y
aromatics and isoprene during each of the sampling periods, are listed
Table 2. Acetaldehyde concentrations were lower (2.9 ppbV) in the *** J*
hours, and gradually increasing to higher levels (6.3 ppbV) in the ***
noon. Similar trends were observed for methacrolein, methyethyl keto^ ^
methylvinyl ketone and butanal. However, for these compounds only *
slight increase in absolute concentrations was observed. Propanal
centrations remained relatively constant throughout the sampling Pefl
Conversely, the concentration trends observed for benzene, toluene an
isoprene demonstrated high levels early in the day, gradually decrea$*g
to lower values in the afternoon. These diurnal concentration pr°^ jjje
are as expected at a receptor site in a polluted atmosphere, such as
Claremont site, where the photochemical reaction products reach maxi"1
in the afternoon hours.
During the last two days (8/28 and 8/29) of field sampling, a CO|BP
rison study was conducted to investigate the carbonyl concentration ^
difference between the two sampling sites, Claremont and Long Beach
ranges of concentrations observed at these sites during those two
are presented in Table 3. The data demonstrate that the concentra
of carbonyls and aroraatics are generally higher at Claremont than at
Long Beach, Since Long Beach is located in a general source area tne^ ^
results support the contention that most of these compounds are fo
the polluted urban atmosphere rather than being a primary emission
local sources.
254
-------
ACKNOWLEDGMENTS
,, Funding for this study was provided by Combustion Engineering, Inc.
p°rporate Technology as part of the Air Toxics Monitoring Program.
^mission to colocate our samplers at the SCAQS sites was provided by
s e ^oject Coordination Staff. Assistance in the analysis of duplicate
aj»ples by GC/MS was provided by Sharon Reiss and Laura Burns of the EMSI
^oratory.
P. Carlier, H. Hannachi and G. Mouvier, "The Chemistry of Carbonyl
Compounds In The Atmosphere - A Review", Atmos. Environ. 20:2079-
2099. (1986).
F.V. Lurmann, A.C. Lloyd and R. Atkinson, "A Chemical Mechanism
for Use in Long-Range Transport/Acid Deposition Computer Modelling",
J. Geophys. Res. 91:10905-10936. (1986).
D. Grosjean and K. Fung, "Collection Efficiencies of Cartridges
and Microimpingers for Sampling of Aldehydes in Air as 2,4-dinitro-
Phenylhydrazones", Anal. Chem. 54:1221-1224. (1982).
4<
D. Pierotti, "Analysis of Trace Oxygenated Hydrocarbons in the
Atmosphere", submitted for publication.
5. n
"• Grosjean, "Formaldehyde and Other Carbonyls in Los Angeles
Ambient Air", ES&T 16:254-262. (1982).
255
-------
OF A CRYOGENIC PRECONCENTRATION TECHNIQUE AND DIRECT INJECTION
TOY THE JGAS CHROMATOGRAPHIC ANALYSIS OF LOW PPB (NMOL/MOL) GAS STANDARDS OF
"C ORRAMTr rnMPnTTNHQ
COMPOUNDS
c*°rge C. Rhoderick
nter for Analytical Chemistry
Cat nal Bureau of Standards
ltltersburg, Maryland
Co There Is an increasing need for multicoraponent gas standards
(p J^ning volatile toxic organic compounds at the low parts-per-bill ion
St i*i nmol/mol) level for use in environmental monitoring programs.
j^ndards containing many organic compounds, both halogenated and non-
°8enated species within the same mixture, can be very difficult to
e at the 1-15 ppb concentration level. Analyses of low level
omponent mixtures have been done using several different techniques.
18 £"romatography employing packed and capillary columns In both the
and the temperature program modes has been used to separate
n s*-mPle an<* complex mixtures. Original work was done using
columns with a flame-ionlzation detector (FID) and large sample
raL> ^or direct Injection. Both packed and capillary columns
used with an electron-capture detector (ECD) to analyze for
Senated compounds at low ppb levels. The analysis of multtcomponent
at -!Jres containing both halogenated and non-halogenated organic compounds
*6n Ppb level requires both FID and ECD detectors because of the lack of
tj S^tivity of the FID to halogenated compounds at these low concentra-
tjK s- The imprecision of replicate injections of a single sample using a
*lv an<* a ^® m^ sample ^-s B°°d for non-halogenated compounds, 1-5% rela-
tmp ' but poor for the halogenated species, 7-25% relative. The
&CD 6c^si°n was excellent for those compounds that are sensitive to the
Sj^1 0.1-0.3% relative. Therefore, to measure all the compounds In a
*tic analysis, a cryogenic preconcentration technique was developed to
teij 6ase the sensitivity of both types of compounds to the FID,
t6c?erature programming was coupled with this cryogenic preconcentration
^ to increase the quality of baseline separations.
259
-------
Introduction
The concern over volatile toxic organic compounds present in the
environment has increased over the past several years. Federal and
agencies have implemented monitoring programs to measure the levels of
certain toxic organic compounds in the air, concentrating on particula*
areas such as workplace environments, hazardous waste dumpsites and
hazardous waste incineration. A number of the compounds under study &*e
considered carcinogenic to both humans and animals, including fish1'2-
There have been many health risk studies done by the U.S. Environmental'
Protection Agency to determine which of these compounds are posing the
greatest risk to human health3. For the past six years, the National
Bureau of Standards has been involved in research and development of gaS ^
standards containing volatile toxic organic compounds at concentrations
10 ppm to as low as 1 ppb*•5•6. Along with the development of these
standards, analytical methods had to be developed for the detection an*.
quantification of these compounds. As the lower concentration ranges o
1-20 ppb were approached, the measurement process became much more com? ^
and difficult. This paper will describe the methods of analysis used &
low ppb mixtures containing a few toxic organic compounds as well &s ®°
complex mixtures containing many compounds. The advantages and
disadvantages of these methods will also be discussed.
Experimental
t- tbe
Gas mixtures of volatile toxic organic compounds were prepared ** e
nominal 15 ppb level by a raicrogravimetrie technique developed at NBS '^j
The organic compounds used for the first standards in this study wef® v
chloride, chloroform, carbon tetrachloride, benzene and tetrachloro-
ethylene. These compounds were selected by EPA comprising Group 1
air monitoring programs. Later, mixtures were prepared containing i? "p
volatile toxic organic compounds with concentrations ranging from 5-1' *
The compounds present in these mixtures are vinyl chloride, bromometti8*1
trichlorofluororaethane, chloroform, carbon tetrachloride, 1,2-dichl°r°*
ethane, 1,1,1-trichloroethane, benzene, 1,2-dichloropropane,
ethylene, 1,2-dibromoe thane, tetrachloroethylene, toluene,
ethylbenzene, styrene and o-xylene.
Several methods were then used to analyze the mixtures. The
method involved the use of a gas chromatograph equipped with a FID **fl'
detector temperature of 250 °C. Nitrogen carrier at a flow rate of 3"
mL/min was purged through the chromatographic column which was"made °*
2.4 m by 3.2 ram OD stainless steel, packed with a 10% loading of P0^"
ethylene glycol modified with nitroterephthalic acid on a 100/120
support and operated at an isothermal temperature of 80 °C. The gas 8
valve contained a 10 mL sample loop.
The second method involved the use of a GC equipped with an BCD •
operated at 350 °C. A wide-bore capillary column of 60 m by 0.75 mm J e
borosilicate glass with a 1.0 micron thick film of dimethyl polysil°**eC
was used at an isothermal temperature of 50 °C. The column flow was *
10 mL/min with a make-up flow of 25 mL/min. The gas sample valve wfl3
fitted with a 0.1 mL sample loop.
The first and second methods were used to analyze the Group 1 ^
mixtures. The 17 component mixture was analyzed using a 2.4 m by 3.*
stainless steel column packed with a 1% loading of polyethylene glyc° fl
modified with nitroterephthalic acid on a 60/80 mesh graphitized cfljr £
black. The column was operated with a temperature program starting a
260
-------
at ^ ^°r 3 Tnlnutes tlien Damped to 220 °C at 8 °C/mIn and the FID was set
250 ec. The nitrogen carrier flow rate was 40 mL/mln and a 10 tnL sample
lo°P vas used,
, The third GC method of analysis used the same column as described
ove for t^e anaiysis Of thB 17 component mixtures. The temperature
Used started at 1W °c for 5-$ minutes then ramped to 220 °C at 25
, -n-te column carrier gas flow rate was 40 mL/mln. of nitrogen and a
sample loop was fitted onto the gas sample valve. The FID operated
teiaPerature of 25° °C. The sample flow of 50 mL/min was maintained by
°^ a raass flov controller. The sample was trapped In the gas sample
*or four minutes by submerging the sample loop in a dewar o£ liquid
At the end °f t*ie traPPinS period, a valve command from the GC
*tit Eas sample valve to the inject mode. The dewar of liquid
fl r^en was imiediately replaced with a beaker of hot water (100 °C) to
"as VaP°rize the trapped organics onto the column. The Group 1 mixture
^ analyzed by this method. The same method was employed to analyze the
j °0mPonent mixture except the column was the same 60 m wide-bore capil-
llti US6d in met^od 2i '^ie temperature program started at 32 &C for 12
frasUt6S> ttien ramPed to 17° °G at 4 "C/min. The sample flow of 50 mL/min
trapped for 5 minutes for a volume of 250 mL.
Sults and Discussion
a£t, The analysis of the Group 1 toxic organic gas mixtures by the first
that described- FID and a 10 mL direct sample injection, resulted in data
tfte Was usable but the imprecision of the analysis was poor. Examples of
t£ , •"'P'tecisloiv the standard deviation of the mean area counts from
c&4 Cat* Injections of a single s-anple, for tha FID procedure, for each
d in a m:Lxture are shown In Table I. The imprecision for vinyl
and benzene, respectively 5% and 1% relative, are very good for
Cotlcentration level. However, the imprecisions for the chloroform,
es ?n tetrachloride and tetrachloroethylene, 7-25% relative, are poor and
Xp *" ^n n*-E^ uncertainties in the assigned concentrations. One might
ct fch« one and two carbon containing compounds to exhibit about the
iresponse to the FID but the chlorine atoms have a quenching effect and
Ce the sensitivity of those compounds to the detector.
Iti^j. achieve better precision of analysis, another method was needed to
liss *ase the sensitivity of the compounds. This second method involved the
It ^ an ECD. The ECD is highly sensitive to halogenated compounds, but
ch^ * essentially non- responsive to benzene and has a low response to vinyl
for e- Problems occur when trying to introduce a large enough sample
CoaiP°unds containing ot\e chlorine or fluorine because the detector can
ejtaj, *r°acin tetrachloride and tetrachloroethylene was excellent,
thfc *ng f10"1 0.15-0,29% relative. So to achieve good im
__ „ ,_ good imprecision for all
°mPounds in the Group 1 mixture, two methods of analysis are needed.
«ie research and development program at UBS proceeded as mixtures and
nards containing up to nine compounds were developed. Both the FID and
BO* vere used to analyze these mixtures. When the 17 toxic organic
3v standard was developed, these two procedures were also applied.
£otwtfr> the complexity of this mixture resulted in a very long run time
Qie * 6 analysis of as much as 40 minutes, A second problem was posed by
act that these mixtures ranged from 5-15 ppb. When the samples were
261
-------
analyzed using the first method, FID and direct Injection of 10 mL,
the compounds did not give a response as can be seen in Table IV. The
imprecision of replicate injections of a single sample ranged from 0.1'*
relative for those compounds which were sensitive to the FID. So to .
measure the other compounds in the mixture the second method was used tfi
an ECD. Once again two methods were required to analyze a low ppb,
mixture, and it required an enormous amount of time.
It became more apparent that there would be more involvement in
ppb mixtures which would be v>-ry complex. So a third method for analysi
was developed which involved temperature programs and cryogenic precoti"
centration of the sample. The Group 1 mixture was analyzed using this
method and the results are illustrated in Table III. The imprecision8
the compounds in this mixture were very good ranging from 1.1-2.4% rel**
tive. Since the results were so good, the cryogenic preconcentration
method was applied to the 17 component mixtures. The results are Sivelltfl(
Table IV which shows that the imprecisions (RSD's) for the compounds **
this method range from 0.2-2.5% relative, with 15 of those species
imprecision of less than 1.0%. These values are considered to be e
at this concentration level, especially when one considers that the
preconcentration of the sample is a source that might affect the
reproducibility of replicate injections.
Conclusions
There are several GC methods of analysis that can be applied to ^t
measurement of volatile toxic organic compounds in gas mixtures. O*16 ,ji
consider the parameters with which the experiment must be performed , s^
as the amount of time to do the analysis and the required precision a°
accuracy. Since the determination of the Group 1 compounds (a rathe* • $
simple mixture) requires a rather short run time (10-15 minutes), no »• \l
which of the three methods is used, then the first two methods would "•
In the lowest analytical imprecisions. The analysis using an ECD wot*1
result in excellent imprecisions, which cannot be equaled using FID & ^
cryogenic preconcentration, for the halogenated compounds. The benZe°
vinyl chloride could be determined by the first method, FID and 10 fflLffl
direct injection, and give good imprecisions. However, if time is *
and precision can be sacrificed for the halogenated species, then tl»e
preconcentration technique is the better procedure and one can
the compounds using one method.
When considering the analysis of a complex multi-component
time becomes the determining factor. It becomes more difficult to W^
the compounds which may increase the run time for a single injectio* y
accomplish the separations. When using the ECD, it can be very diff* $>
to use temperature programming due to the sensitivity of the detecto*
even very small amounts of column bleed when ramping the temperature-
Column bleed is hard to control even when doing a column compensati" '
Therefore, better results are obtained using an isothermal column te ft fi
ture, which for this particular multi-component mixture, resulted -^^
minute analysis per injection. The run time using temperature ytogf f
and a FID was 25 minutes for direct sample injection and 40 minutes ,
sample preconcentration method. It would take less time to do the *
for all the compounds if the sample preconcentration method is used
than using both FID and ECD. The data show that the preconcentrati°
technique is definitely better than using a direct injection of 1° ^
sample, especially for the halogenated species.
262
-------
In summary, the three methods described are all feasible for the
d atysls of toxic organic gas mixtures at the low ppb level. Factors that
th erm*ne which method is the most appropriate include the amount of time
t,at fche analyst has to complete the work, the complexity of the mixture,
- type of organic compounds present in the mixture and the level of
alytical precision that is required.
AcWledgment
0_ The author wishes to acknowledge Darryl von Lehmden and Howard Crist
Sy e U-S. Environmental Protection Agency's Environmental Monitoring
, ems Laboratory for their support of this work. This work was supported
under Interagency Agreement DW-13932187-01-0 with the U.S. EPA.
feren
Monographs on the Evaluation of Carcinogenic Risk of Chemicals to
International Agency for Research on Cancer, World Health
°rganization, Geneva, Switzerland, 1972-present.
5' C* Malins, B. B. McCain, D. W. Brown, M. S. Myers, M. M. Krahn,
5--L. Chan, Environ. Sci. Techno1. 21: 765-770 (1987).
1 «J. von Lehmden, Conference on Recent Developments in Monitoring
^SShfids for Tnxina jn the Atmosphere. Boulder, CO, July 1987.
4
W- P. Schmidt, H. L. Rook, Anal. Chem. 55: 290-294 (1983).
5^
• c- Rhoderick, W. F. Cuthrell, W. L. Zielinski, Jr., Transactions
j*££A/ASQr Specialty Conference on Quality Assurance in Air Pollution
its. T. R. Johnson, S. J. Penkala, Ed.; APCA, Pittsburgh,
(1985).
6,
• c- Rhoderick, W. L. Zielinski, Jr., Conference on Recent
y*^Slppmar>r.s jn Monitoring Methods for Toxics in the Atmosphere.
°ulder, CO, July 1987.
I- Imprecision data using method 1, FID and 10 mL direct injection,
for 15-20 ppb Group 1 organic gas mixture.
Carbon Tetrachloro-
Chloroform Tetrachloride Benzene ethylene
157 1020 444
169 1013 475
133 998 447
156 998 512
165 1001 500
193 1012 505
14^ _994 41i
160 1005 474
19 10 32
12% 1% 7%
values represent the analytical imprecision.
263
-------
Table II. Imprecision data using method 2, ECD and 0.1 mL sample, £°r
15-20 ppb Group 1 organic gas mixture.
Vinyl
Chloride
Chloroform
Carbon
Tetrachloride
Tetrac
bio'0'
nr
avg =
sd =
rsd9 =
Benzene
nr
'These percent values represent the analytical imprecision.
Table III. Imprecision data using method 3, FID and cryogenic
preconcentration, for 15-20 ppb Group 1 organic gas
g
^11
Vinyl
Chloride
avg =
sd =
rsda -
63.69
63.02
64.85
63.85
0.93
1.5%
Chloroform
12.60
12.60
12.86
12.69
0.15
1.2%
Carbon
Tetrachloride
10.91
11.05
11.43
11.13
0.27
2.4%
Benzene
115.73
115.38
118.22
116.44
1.55
1.3%
'These percent values represent the analytical imprecision.
Table IV.
Vinyl chloride
Bromomethane
Chloroform
Benzene
Toluene
Chlorobenzene
Ethylbenzene
Styrene
o-Xylene
Comparison of imprecisions for direct injection and
preconcentration using FID for 17 component organic
mixture at 10 ppb.
Direct
(lOmL)
3.6%
3.2%
6.8%
0.6%
1.6%
3.5%
3.2%
1.4%
3.1%
Precon.
(250mL)
0.4%
1.9%
0.8%
0.4%
0.4%
0.5%
0.3%
0.3%
0.2%
Trichlorofluoromethane
1,2-Dichloroethane
1,1,1-Trichloroethane
1,2-Dichloroprppane
Trichloroethylene
1,2-Dibromoethane
Tetrachloroethylene
Carbon tetrachloride
Direct
(lOmL)
264
-------
Cnl2MATED ANALYSIS OF MULTICOMPONENT
PRESSED GAS MIXTURES CONTAINING
PER BILLION CONCENTRATION OF
ORGANIC COMPOUNDS
B. Howe,
Jayanty
R search Triangle Institute
ie*rch Triangle Park, North Carolina
U. W J. Von Lehmden
Res*' ^nvtronmental Protection Agency
Triangle Park, North Carolina
Jto automated cryogenic preconcentratlon system coupled with capillary
gas chromatography has been developed for analyzing compressed gas
"s containing trace levels of toxic organic compounds. These
"s are made available to federal, state, and local air pollution
agency personnel and their contractors for conducting performance
toea*,1"5 Curing hazardous waste trial burn tests and ambient air
c°ntr i
*U°
automate^ system employs a multlpositlon rotary valve for cylinder
' a Nutec^ cyrogenlc trapping/cryogenic focusing system, a
- Packard 5880A gas chromatograph with both flame lonlzatlon and
s°ftw n caPture detectors, and a personal computer with process control
"1"6 to ena^^e unattended analysis of up to eight cylinders. The
nded" feature allows for automated analyses after normal working
thus Increasing significantly the 24 hour output.
calibration of the detector response and analytical quality
°f *° 1s performed with cylinder mixtures prepared by the National Bureau
^dards (NBS) for the U.S. Environmental Protection Agency under an
Agency agreement.
265
-------
Introduction
Under contract to the U.S. Environmental Protection Agency, The
Research Triangle Institute conducts a program in which compressed gas
cylinders containing ppb levels of toxic organic compounds are made ^
available to federal, state, and local agencies, and their contractors,
performance audits during hazardous waste trial burn tests and ambient
measurements.
Currently five different groups of cylinders are available with eaC
group consisting of cylinders containing a specific mixture of organic
compounds diluted in nitrogen. Cylinder mixtures are available in . in
concentrations ranging from 1 to 10,000 ppb. A listing of the compound
each group and concentrations available is shown in Table I.
Each group of cylinders was prepared by a commercial manufacturer^
also provided a certified quantitative analysis of each cylinder mixtui •
Upon receipt of a new cylinder from the manufacturer, RTI performs an
analysis to check the manufacturer's certified value. Each cylinder 1*
subsequently analyzed after 2 months, 6 months, and 12 months, and thep „{
every year thereafter to determine the long-term stability of the comp°
concentrations. Stability data on all compounds shown in Table I are
periodically published.1
$
Quality assurance is provided by EPA auditing of RTI using NBS nm
standards prepared by the National Bureau of Standards under an
interagency agreement. The NBS standards are subsequently used by
calibration of the detector response and as a quality control check
analysis of the cylinders obtained from the commercial gas supplier.
In 1983 when this program started, RTI used an analytical procedj^
for stability study analyses involving packed column gas chromatograpW
with either flame ionization or electron capture detection.2 Gas samp
volumes from 1 to 10 cm3 were injected directly on-column and detectoj"
responses near their practical lower limits were used for quantitatiflS
cylinder component concentrations.
t.etf
For recent analysis, an automated cryogenic preconcentration SV «t0
has been implemented which allows injection of larger analyte masses
fused silica capillary column with simultaneous detection by flame ^
ionization and electron capture detectors. The automated analysis sp
configuration, analytical procedure, and some analysis results will "
presented in this paper.
System Description and Analysis Procedure
J \<
A block diagram of the automated analysis system recently deV
shown in Figure 1. System components include a nine-port multipos
valve which is electrically actuated for selecting the appropriate c.
for analysis, cylinder gas pressure and mass flow control devices, a
cyrogenic trapping unit, a personal computer with monochrome monitor
266
-------
5880A gas chromatograph with both flame 1on1zat1on and
caPture detectors, and GC peak data processing with the 5880A
4 terminal and an HP 3393A computing integrator.
anal Up- to ei9nt cylinders can be connected to the system for automated
sta H s* TyPica^y the eight cylinders consist of an NBS quality control
anal ' d n^trogen cylinder for purging, and six audit cylinders to be
f^Jyzed. Each cylinder is connected to the nine-port valve with a CGA-350
stai ' a ^ow snubber, a check valve, and a length of 1/16 in. OD
Ss steel tubing. The flow snubber serves as a safety device to
^P^ decompression of the cylinder in the event of a break 1n the
9 downstream. The check valve prevents cross-flow from other
during multiport valve switching or cross-port leakage in the
the ort Va1ve- Tne nine-port valve actuator is computer controlled and
p0 ,^alve position can be selected at random by inputting the valve
•tion in binary coded decimal format to the actuator.
Is u ter all cylinders for an analysis are connected, a purging procedure
fitti6^ to ensure removal of any residual room air from the connecting
Port and tubing. Tn1s 1s accomplished by manually advancing the nine-
ty Va^ve to select the first cylinder for purging. The main cylinder
6 *S t'1en °Penec' momentarily to pressurize the connecting fittings and
9« A three-way valve located after the multiport valve is switched to
s depressurize the system to ambient pressure. This process of
1S flzati on/venting is repeated 3 additional times. The nine-port valve
ac*vanced to select the next cylinder and the process 1s repeated
cylinder connections have been purged.
not ^n t^ie vent Pos^t1on, the outlet of the three-way valve
to a S'in9^e staQe high purity regulator which is set at 30 pslg
ressure to Prov1c'e the appropriate Inlet pressure to a mass flow
r* T^e mass ^^ow contr°ner 1s set at 20 cm3 per minute for
ing gas flow Into the cyrogenic trapping unit.
BA$J ,/^ter cylinder connection and purging is complete, the operator runs a
4ncj l;hcoi):iputer program for controlling the number of analyses to perform
v secluence ^n which cylinders will be analyzed. Any number of
,!s?s can be performed, but only a maximum of 8 different cylinders can
typical scheme for performing analyses over a 24 hour period 1s
Table II. Following this scheme, 9 audit cylinders can be
^n a 2/^ nour peHod with the NBS quality control standard being
twice during the 24 hour period.
the operator verifies that the correct sequence of analysis has
the system then is controlled entirely by the personal
Shown in Figure 2 1s a schematic of the automated analysis
S control. The personal computer is configured with two hardware
for controlling the cryogenic trapping unit. These are an
analog to digital converter and a 24 bit parallel digital
267
-------
Input/output Interface. Control of the cyrogenlc trapping process 1s
achieved with a software package called Notebook from Laboratory -
Technologies Corporation. This software enables digital control of son
state relays contained in the cyrogenlc trapping unit for cryogenic
solenoid activation, six-port valve switching, and automatic start of &
data acquisition.
An additional digital Input/output board which 1s programmed 1n B^5
1s used to control the electric valve actuator for the nine-port rnuHl-
pos1t1on valve.
When the process control routine 1s begun, the cryotrap cooling
solenoid Is opened which allows liquid nitrogen to flow to the cryotrap'
The trap temperature 1s monitored by a thermocouple and the amplified ^
thermocouple signal 1s used as a trigger for beginning sample gas traP?
When the cryotrap temperature reaches -150°C, two six-port valves
the cryotrapplng unit are switched to direct sample gas Into the cr
After 10 minutes of trapping at 20 cm3 per minute, a valve switch 1s
performed to divert the sample gas to a vent. After 10 seconds, the
nitrogen flow to the cryotrap 1s ceased and the trap begins heating
150°C. After 20 seconds of heating, another valve switch diverts heU" r
carrier gas through the cryotrap. From the cryotrapplng unit, the can
gas with the analyte compounds passes through a heated 1/16 1n. OD
stainless steel tube to the capillary column Inlet.
igfiV*
The first portion of the capillary column is housed 1n a Teflon *• ^
through which liquid nitrogen 1s passed during sample injection to acj ^
cyrofocuslng of the analytes. The cryofocuser is kept cooled to -I0?n»c t0
two minutes after the injection 1s made and then heated rapidly to W r$
begin chromatographing the analytes. The GC column consists of 50 mel •
of 0.32 mm ID fused silica which is wall coated with 1.2 microns of c ftf
dimethyl polysiloxane. The column temperature program consists of 35
4 minutes, 6°C per minute to 90°C, and 15*C per minute to 170°C.
t.ftf
The carrier gas outlet flow is split approximately 10:1 between *
FID and ECD through the use of an SGE variable split outlet splitter. .^
Nitrogen make-up gas is added prior to the splitter to prevent dead v°
effects.
An example chroma togram of a Group 5 mixture {18 components in
nitrogen) is shown In Figure 3. The FID signal 1s used for quant
all components except Freon 11, which 1s quantitated by ECD. The
signal is processed with the HP-5880A Level 4 processor and the EC"
is processed with an HP-3393A computing Integrator.
Quality Control Results A
to "
As discussed earlier, NBS standards for each group are
for both calibration of the detector response and for quality c011*1^) to .
during routine analysis of commercial mixtures. One standard Is us '
establish the Instrument response factors prior to initiating anal
268
-------
cylinder group. The other standard is analyzed with each set of
ers attached to the system for automated analysis. If the measured
I°ncentrat1on of any component 1n the NBS quality control cylinder differs
y "lore than 10 percent from the assigned concentration, the Instrument 1s
ca'1brated and the sample cylinders are reanalyzed.
tahi control test results for Groups 1, 2, and 5 are shown 1n
con IIIf Iv and Vt Both NBS concentrations and the RTI measured
4y,"Centrat1ons with their respective concentration uncertainty estimates
the D^Ven* Tne total uncertainties at the 95 percent confidence level for
measured concentrations were calculated using equation 1.
ua = 2((a/2)2 + b? + C2)l/2 (1)
Where Ua = Total uncertainty at 95% confidence level, percent
a = Uncertainty at 95% confidence level for NBS traceable
standard component, percent
b = Uncertainty of calibration standard component response,
percent RSO
c = Uncertainty of sample component response, percent RSD
a^ °f t*ie 9rouPs analyzed thus far with the automated system, the
MR|*ncerta1nty of component concentrations 1s only slightly higher than
si s*andard concentration uncertainty. This reflects the good
e ]on in analyses performed with the automated system. In addition,
1s good agreement between the NBS traceable concentrations and the
ed concentrations for the QC cylinders with less than 5 percent
for many comP°unds' and less tnan 10 percent difference for most
m °.u9h development of a personal computer based automated analytical
1 Cry°sen1c preconcentratlon of analytes, several Improvements 1n
°^ aud^t cylinders have been achieved. These Include: an
ef^c1ency with the ability to analyze as many as nine sample
n a 24 nour Perlod (1n addition to two quality control
1mProved precision resulting from Injection of larger analyte
. greater detector response; significant reduction 1n analyst
vH , lower concentration uncertainties resulting from Improved
nical precision.
theory- th1s system should also be feasible for determining toxic
;oniPound concentrations 1n ambient air samples which have been
n 1ow Pressure canisters. Such an application of the automated
system 1s planned for future work.
269
-------
References
1. Jayanty, R. K. M., C. K. Sokol, and C. E. Decker, "Status Report *J '
Stability of Parts-Per-Billion Hazardous Organic Cylinder Gases and
Performance Audit Results of Source Test and Ambient Air Measurer^11
Systems." EPA Final Report, January 1988.
2. Jayarty, R. K. M., S. W. Cooper, J. Sokash, and C. E. Decker,
Report #1 - Stability of Parts-Per-Billion Hazardous Organic Cy
Gases and Performance Audit Results of Source Test and Ambient Ail"
Measurement Systems." EPA Final Report, January 1985.
270
-------
Table I. Gas Mixture Components and Concentrations
Group 1: Carbon tetrachlorlde; Chloroform; Tetrachloroethylene; Vinyl
chloride: Benzene. Concentration ranges: 7-90 ppb, 90-430 ppb, 430-10,000
ppb.
Group 2: Trichloroethylene; 1,2-Dichloroethane; 1,2-Dibromoethane; Freon-
12; Freon-11; Bromomethane; Methyl ethyl ketone; 1,1,1-Trichloroethane;
Acetonitrile. Concentration ranges: 7-90 ppb, 90-430 ppb.
Groups: Vinylidene chloride; Freon-113; Freon-114; Acetone; 1,4-Dioxane;
Toluene; Chlorobenzene. Concentration ranges: 7-90 ppb, 90-430 ppb.
Group 4: Acrylonitrile; l,3-Butad1ene; Ethylene oxide; Methylene
chloride; Propylene oxide; Ortho-xylene. Concentration ranges: 7-90 ppb,
90-430 ppb.
Group 5: Vinyl chloride; bromomethane; Freon-11; Dichloromethane;
Chloroform; 1,2-Dichloroethane; 1,1,1-Trlchloroethane; Carbon
tetrachlorlde; Trichloroethylene; 1,2-Dibromoethane; Tetrachloroethylene;
Chlorobenzene; l,2-D1chloropropane; Benzene; Toluene; Ortho-xylene; Ethyl
benzene; 1,3-Butadiene. Concentration range: 1-40 ppb.
Table II. 24 Hour Analysis Scheme
Event Number Description
1 Analyze three audit cylinders.
2 Analyze NBS standard for
quality control
3 Analyze six audit cylinders.
4 Analyze NBS standard for
quality control.
Notes: The GC response 1s calibrated with a 2nd NBS standard on an
earlier day during normal working hours.
Events 1 and 2 are performed during normal working hours.
Events 3 and 4 are performed after normal working hours with
no operator present.
Each cylinder analysis consists of three Injections.
271
-------
NJ
Cylinders
9-Port
Valve
Pressure/
Flow
Control
Monitor
t
Personal
Computer
Cryogenic
Trapping
Unit
GC Data
Processors
t t
FID
ECD
GC
-------
Thermocouple
crr*ocouple
Notebook
A/D
Converter
_c
i
1
c
J
I 1 ICI 1 1 IUlsUU|JIC
Amplifier
— -J
0)
cO k^
*«•*• Q^
°§
Is
W CD
"5 -g,
COjf
•r>^ y v ^*«^
BASIC
3
O)
Q
I/O Interface
9-Port Valve
Actuator
\
Data
Acquisttion
Start
PC
Cryofocuser
Cooling
Figure 2. Automated analysis process control.
275
-------
METHOD FOR ANALYSIS OF VOCs
VALLEY AMBIENT AIR
[) oar-raa and R, Eggleton
C?rtlnerit t:'t Chemistry
ln ' v'rginia State Co! lege
' WV 25112
v A method has been developed for simultaneous analysis of several
KO|tlle organic compounds (VOCs). VOCs (B.P. -4°C to 150°C) are
on active charcoal, desorbed with o-dichlorobenzene, and
ln ^e liqujcl extracts and in head-space vapors with dual
'lc|, CdP'llary column gas chromatography . Compounds analyzed
acrylonitri le (ACRN), 1,3- butadiene (BUTD), ethylene oxide
methyl -t-butyl ether cMTBE), and propylene oxide (PRO).
if ication of compounds was made from relative retention
°n aifferent columns recorded in FIDs, and by spiking. ETO. PRO
iw ^ Were confirmed by their reaction with hydrogen chloride and
*"atogram of the derivatives formed.
in t, BUTD, ETO and ACRN are compounds of considerable health concern
Of BUT |('andwhd Valley, There is no EPA approved method for analysis
is a r° and ETO in ambient air. PRO is a suspected carcinogen. MTBE
additive in certain brands of premium unleaded gasoline.
ntal data for ACRN. BUTD, ETO, MTBE, and PRO in the Kanawha
are currently on record.
277
-------
INTRODUCTION
The Kanawha Valley represents a unique combination of
geographical setting, widely variable me tero logical conditions, and
large concentration of industrial facilities, with residential
population in close proximity. Kanawha Valley Toxics Screening Study
(KVTSS)1 was undertaken by the U.S. Environmental Protection Agency
(EPA), in cooperation with West Virginia state environmental agencies.
The major objective of the study was to assess potential long-term
health risks from a selected number of low level non-criteria air
toxics, routinely emitted in the Kanawha Valley.
Air emissions of several hundred chemicals from Kanawha Valley
chemical industry are listed2 for 1984. Several pollutants are
also emitted by county-wide sources3. KVTSS selected twenty
pollutants, that have received a unit risk factor from EPA's
Carciniogen Assessment Group, 1986, to screen their relative risks to
the Kanawha Valley residents. Acrylonitri le, ethylene oxide,
1,3-butadiene and chloroform were the four compounds estimated in
KVTSS to pose upper bound individual cancer risks of over 1 in 1000.
EPA methodology for estimating individual lifetime cancer
from low level routine air pollutants is computed from the products
(1) ambient concentrations, C, (2) exposure constants, E, and (3>
potency, P. KVTSS used modeled data from annual emissions reported
for 1984 for C. Differences in time of residence and quality of
indoor and outdoor ambient air were not included in E. The least
precise parameter in assessment of cancer risks is In the values
assigned to P. It is, therefore, important to improve the most
reliable component, C, in the chain of risk assessment process. No
EPA approved methods are currently available for BUTD and ETO^. No
experimental data on ACRN, BUTD, ETO and PRO for Kanawha Valley are
available at present. Thus uncertain, modeled rather than measured
values have been used in risk assessment for air toxics of major
concern in the Kanawha Valley.
EXPERIMENTAL METHOD
1,2-Dichlorobenzene (Baker analyzed Reagent, HPLC grade) with
added n-octane (24 mg/liter), ODCB, was employed throughout for
preparing standards, chemical desorption of VOCs and wash liquid f°r
the syringe before sample introduction. All liquids analyzed
contained greater than 99% 1,2-dichlorobenzene, approximately 0.6%
1,4-dichlorobenzene (main impurity in the Baker reagent) and 24
mg/hter n-octane as internal standard. Primary standards were
278
-------
by adding calculated volumes of compounds (99+% purity) to
Secondary standards were made by serial dilution of primary
(wt/wt) with ODCB.
A Hewlett-Packard 5880A level IV gas chromatograph equipped with
[ Hewlwtt-Packard, 25 meter, 0.31 mm ID, 1.05 micron film) and
(J & W. Scientific, 30 meter, 0.32 mm ID, 1.0 micron film)
oet ---iV columns were used for separation. Flame ionization
-. ectors were used for quantitat ion. HP 7673A automatic sampler was
! for sample introduction, where ODCB was employed as wash
3y Most of the developmental work on the method was done with
standards and on liquid samples on HP 5880A (rather than
-space vapors on Perk in Elmer 8500/HS-6).
COI Universal constant-flow (SKC and Gillian) pumps were used to
Ct VOCs from amD'ent air on two absorption tubes (400/200 mg
„„. ^t-base active charcoal : SKC 8 226-09) on a dual adjustable flow
for p CSKC * 224-26-02) at different rates (ca 125 and 250 mL/min)
8toD hours. These tubes were capped, placed inside another
ChJ.Perecl test tuoe and refrigerated immediately after collection.
fr- c°dl from the absorption (400 mg> and control (200 mg) segments
^ each sampling tube were placed in separate vials, followed by 2
ljnL OCDB, respectively, and refrigerated until analysis. Samples
•asurable backgrounds in OCDB extract in corresponding control
: were rejected.
'TS AND DISCUSSION
CQC Active charcoal is the most efficient solid absorbent for VOCs.
o base charcoal
-------
Peak area (relative to that of n-octane) to concentration of
VOCs in standards were linear over a considerable working range,
DB-1701 column of medium polarity effected better separation of
several VOCs, but coeluted ethylbenzene and chlorobenzene.
Heavy contamination from particulates, and chemicals at the
science facility of West Virginia State College, currently under
extensive renovation and construction work, has seriously interfered
with our analytical laboratory and quality control of ambient air
analysis in 1988, No suitable space could be found to house the
environmental analysis laboratory, A small room, just enough for the
Hewlett Packard GC system, was provided in another building on the
campus in April, 1988. Discontinuation of EPA funding for 87-88
compelled the lab technician (Rita Eggleton) to quit the program for
another job. The work is, therefore, being carried out at a slow
pace.
The following VOCs were detected and analyzed in the ambient air
at Institute : BUTD, ETO, PRO, ACRN, MTBE, methylene chloride,
chloroform, benzene, 1,2-dichloroethane, trichloroethylene, toluene,
and methyl-isobutyl ketone. Details of ambient air monitoring in the
Kanawha Valley will be communicated later In a separate paper.
Typical chromatograms of VOCs in standards are given in figure I
and the data are summarized in table 1. Figure II shows chromatograrns
before and after reaction with HC1 - demonstrating elimination of ETCU
PRO and MTBE peaks and formation new ones for the derivatives formed-
CONCLUSION
The method described is capable of providing analytical data on
twenty VOCs, including compounds of major health concern in the
Kanawha Valley ambient air, for which preexisting data are not
available. The procedure is rapid, inexpensive, and applicable to
both Indoor and outdoor ambient air monitoring. Preparation of
standards in ODCB is easier than gas standards. Shelf life of these
standards was also found to be satisfactory.
Positive identification for ETO, MTBE and PRO has been effected
by chemical derivatization. The limitations of Tenax GC in collecting
more volatile VOCs and artifacts formation have been overcome by use
of more efficient, low-cost and commercially available disposable
active charcoal tubes. High sensitivity attainable by thermal
desorption and cryofocuslng, employed in GC/MS analysis, had to be
sacrificed in dilution accompanying chemical desorption. This is ,
partially offset by higher absorption capacity of active charcoal, afl
optimization by repeat analysis for each collected sample.
280
-------
Large solvent fronts associated in extracts with more volatile
°'vents are avoided by the use of ODCB, that elutes after VQVs of
Merest, ODCB extracts also allow effective head space analysis of
^ volatile VOCs. Collection of larger sample, and/or over a longer
lo should not pose any serious problem, as we are looking for
^9-term health effect rather than checking TLV-ceilings in
austrial setting or monitoring acute risks from accidental release
tQxlc chemicals.
ES
gp. I- Kanawha Valley Toxics Screening Study : Final Report, U.S.
ft' July 1987
fv, 2. West Virginia Air Pollution Control Commission, 1984
Ssion Inventory, Kanawha Valley Chemical Industry, 1987
E&A 3- Regulatory Integration Division, Office of Policy Analysis,
*' 1987
C0to& 4' Compendium of methods for the determination of toxic organic
SeDTOUncls ln ancient air, EPA-600/4-84-042, R. M. Riggln et al ,
1096
Ckl
T on " work was 3UPP°rtecl in Part °y Environmental Protection Agency
r;JO 1604-01-1, 1986-1987) and by a grant from the National Institute
Studies, Charleston, WV (1987-1988)
281
-------
I'.' It.
•i r
Figure I
Standard VOCs on DB-1
x peaks remoned by HC1
^^
OJ
CO
H
fl
^•2
• vo
Ol I—
rH OJ
^^f
ON f
^ I-i;'j G
'"'I * ^t*
.1. llj -^
r v- *
n i
•
A_
ro
-j
-j
i-
^
CO
rH
in ro
rH VD
'ON C\!
•— • IT\ rH
H O-.
S
c% oj ON vo t—
***^ rH "^ ^* rH rH
O *— ^ l**^ **^ """"
rH t— H ON t— ..
^, ^- J — t— OJ [
t- . — .l/\ rH-VD •• "
^,CO r~. rH I— ^T ONO
O --- O — ' f>- rH t--
rH ON r-, H h- CO. (IT
• l^ " •— • VO |. .
-» • ' rH • '.ft,
-» ' O\ V^ ' >
Cl "^ ^ '- ^' o, " oi
H • f - ''•! •
,; "i
-1
^
•' ^ i-
. 'r'
IT)
L.
_
•
" '".-'
™
o\
_
ON
-5
d
L_
f .
• •
, — -^
l_ '^ — ^*
in o*
^. ITl
M 1
Tt
_ iw m
f^ • £ «•*
'^ (VJ ^1 .
-• -f rv
J
llM
•
IT. ,f,
ll>
00
rt
ll")
ir
D
9
a
-
• «3
^ .
* 0
— 1
-
STANDARD + HC1
o new peaks at
3.0lt, l*. 79, 5-55.
from derivatives
ETO, PRO and
°f
282
-------
In
r-
_Figure 11
Standard VOCs on DB-1701
x peaks removed by HC1
-a-
f-t
irv
IM
in
CO
in
•
E f>~°
in
ro
•
T'J
to
•
*
in
K
U
in
w Q>
-t ~~~
— VD
OJ
STAIJDARD + HC1
x new peaks at k. Ol*,
8.95, 9.26, 9.90
from derivatives of
ETO, PRO and MTBE
283
-------
Table 1
peak RT mi n
tt < i \ gure I )
VOCs
Bu tad i ene 1,3
Eth/1ene ox i de
Propylene ox i de
AeryI on i tr i1e
Me thy!ene chloride
Methyl-t-butyl ether
Ch 1 oro-f orm
Di chloroethane
Benzene
Tr i chloroe t hy1ene
Methyl-i-butyl I
Toluene
n-Octane
Chlorobenzene
Ethyl benzene
m,p-Xylene
o-Xylene
SOLVENT
Di chlorobenzene
Di chlorobenzene
de
ther
,2
e
e tone
(1 ,4)
(1 ,2)
ca
ng
7
14
14
7
40
7
70
17
7
20
7
5
5
7
7
5
5
'A
>0.6
>99
DB-1
0.31
: 25 m, 1
mm ID
Standard +
1
2
3
4
5
6
7
3
9
10
11
12
13
14
15
16
17
18
19
1 .79
1 .86 <3
2.36 (4
2.51
2.76
3.44 <5
5
4. 10
4.59
5.10
5.91
6.67
7.47
8.46
9.17
9.59
9.79
10.27
12.63
13.49
.05
HC1
.04)
.79)
.55,
.98)
peak RT min
* (figure HJ
DB-1701: 3° * ip
1.0 u, 0.32^
Standard f
1
2
3
6
5
4
7
10
8
9
13
12
1 1
14
14
15
16
17
18
2
2
40
.80
3.55
5.01
4-32
4.25
6
7
6
6
9
9
8
11
11
11
,15
,28
,57
.96
,10
,00
,50
,11
,11
,27
11.92
14.67
15.55
284
-------
of a Tekmar 5000 Thennal Desorber
with a Hewlett-Backani GC/MSD for
of Volatile Organic Compounds
on Tenax from Ambient Air
A- LaRue and Linda R. Berrafato
ty of Medicine and Dentistry of New Jersey
Wood Johnson Medical School
e
MJ Q8854
is presented of new instrumentation for the
ive analysis by capillary GC/MSD of selected VOCs collected
Lent air onto Tenax. A Tekmar 5000 is used for thermal
10^°^0f actively collected samples with total sample volumes in
L range. After desorption, the sample is automatically
to a Hewlett-Packard 5890 GC/5970 MSD for separation and
rV A Hewlett-Packard HP-1 50 meter, 0.2 mm ID column with 0.5
•hyl silicone gum coating used with helium carrier gas and
v ^f J^^aMnircr separates the compounds. The analyte is then
" with no splitting into the MSD. Due to vacuum
IB MSD, a plumbing modification is necessary in the
QfU^ ^ fnsure consistent operating conditions by reducing the
>om air introduced into the system with each sample.
- of fourteen compounds are analyzed, and results are
srravvUng dynamic ranges, selected calibration curves,
of standards, and R2 values.
285
-------
Introduction
Analysis of air samples for volatile organic compounds (VOCs)
has become increasingly important to environmental monitoring, and
a variety of equipment is now available for automated analysis of
such samples. Some instruments have been used for quite some time
for other types of analyses, while certain instruments are
relatively new to the marketplace. The interfacing of two or more
of these instruments can improve sample throughput and
reproducibility if the instruments are appropriate and compatible.
An evaluation is presented of a system using the Tekmar 5000
thermal desorber with the Hewlett-Packard 5890 gas chromatograph
(GC) / 5970 mass selective detector (USD) for the analysis of VOCs
collected on Tenax from ambient air.
Methods
Ambient VOCs for this analysis were actively collected using
stainless steel traps (5/8" diameter, 7" length) containing 1.2
grams of Tenax. Approximately 20 liters of air was sampled at a
rate of 5-15 ml/minute over a period of 24 hr. Samplers designed
and constructed in this lab protect the traps from external
contamination during sampling collection.
Thermal desorption of the Tenax was accomplished in the
Tekmar model 5000 by placing a trap into the air-tight furnace and
initiating the "Run" program. Room air was purged from the trap by
a flow of helium carrier gas, after which the furnace was heated
for desorption and subsequent sample deposition into a cryogenic
trap (Cryo Trap-1) cooled by liquid nitrogen. After desorption an
eight-port valve automatically switched, Cryo Trap-1 was quickly
heated, and the sample moved through a heated transfer line to a
second cryogenic trap (Cryo Trap-2) located at the head of a 1
meter pre-column of deactivated fused silica capillary tubing.
Cryo Trap-2 was ballistically heated upon completion of sample
transfer, volatilizing the sample and allowing an injection to be
made rapidly onto the pre-column. An interface cable supplied by
Tekmar automatically sends a "Start" signal to the MSD Chemstation
computer to initiate the temperature program and data collection.
Because of what we perceived to be excess plumbing, the unheated
transfer line was attached directly to the tee at Cryo Trap-2. We
also disconnected the injection port of the GC from the system,
thus eliminating a length of unheated tubing.
The sample was separated and detected using the following
GC/MSD conditions:
* An HP-1 methyl silicone gum fused silica capillary
column, 50 m length, 0.2 mm id, 0.5 micron coating
* 99.9999% helium carrier gas
* 32 cm/second linear velocity
* Initial column temperature 30°C, 7.00 min,.
4°C/min to 75°C for 8.00 min.
70°C/min to 200°C for 2.00 min.
286
-------
* Splitless injection
* Capillary direct (splitless) GC/MSD interface
* 200°C interface temperature
* Ion range 45-170 M/Z collected from 3.00 - 30.04 rain
j, The Chemstation is equipped with optional sequencing software
^ aHow automatic detection, integration and report generation.
a^ ard copy of the report was sent to the printer, and the data was
Permanently stored on magnetic media for future reference.
Modifications
CQ . ^w° modifications are necessary to insure consistent
m0j :-toiis in the system. A purge/desorb gas vent located on the
lar 5000 must be manually opened and closed in order to prevent
d0w°6 am°unts of room air from entering the system during cool-
cap . the cryogenic traps. A standard 1/16" compression fitting
Vacu S suitable for this purpose. The MSD has a relatively small
pumping system and hence it is advisable to reduce the
°f excess gas injected per run. The sensitivity of the MSD
D affected by repeated injections of room air, with a marked
°r W n resP°nse after the injection of a large amount of room air
Cry0 T61"' ^e manual closing of this vent during cooling of the
c°lum raps maintains more stable conditions within the analytical
n and MSD, thus insuring better reproducibility.
Secondly, thermal insulation of at least two zones of the
desorber have proven helpful for more efficient sample
A section of tubing connecting Cryo Trap-1 to the heated
ne is not heated at the point where it is joined to the
"carriPort valve, and wrapping this line (together with the
QUter r~*-n" line) with thin strips of aluminum flashing and an
*6cOrid 6Ve °f Slass wo°l keeps the volatilized sample from
enning in the steel line. The unheated transfer line is
wrapped to the heated transfer line with aluminum strips
wool, thus preventing sample condensation in the unheated
Uring deposition onto Cryo Trap-2.
TK
^° Tr 6 e^fect of closing the purge vent during cool down of the
« °mahrl»^ Can ^e seen by comparing the total and extracted ion
found in Figure 1. There is a marked decrease in
th ° unretained, low molecular weight compounds detected
^ v^nt is closed. This decrease is extremely desirable
e an ^Ontributes to more stable conditions in the transfer
nalytical column and MSD.
S ^°r dynamic ranges and correlation coefficients for
C
ttltt is°"JPOUnds a" listed in Table I. In general, the upper
fctns a deterrained by the point at which the slope of the curve
e^iap0 neSative deviation. The calibration curves have been
]re dftt ed to be more sensitive at the lower end by including
Points of lower concentration. This was done to more
287
-------
accurately quantify ambient samples which, in general, are of lower
concentration and fall at the low end of the curves. Curves are
presented in Figure 2 for 1) a low molecular weight compound,
methylene chloride, 2) a chlorinated compound, trichloroethylene,
and 3) an aromatic compound, benzene,
Reproducibility of standard injections from a gas dilution
bottle reflect the change in sensitivity and chemical environment
in the transfer line, analytical column and MSD. Results for
daily standards for a one week period in December, 1987, during
which at least seven samples per day were run indicate a range of
+ 180% of the calculated amount. As the system has been used, the
reproducibilty has improved to + 40%. Due to the method of sample
introduction from the gas dilution bottle, the best one can expect
for reproducibility is + 30%. Care must be taken to insure
integrity of the standard. Mininert valves and static dilution
bottles supplied by Tekmar need to be checked for deviations in the
cap liner and bottle neck.
Discussion
The above described system has several advantages for routine
sample analysis where automated operation and high sample
throughput are desirable. Because the thermal desorption process
is controlled by a preprogrammed method, timing and temperatures
are very reproducible from run to run, and the system is extremely
easy to use. Parts, supplies and service for the entire system are
readily available, however the two manufacturers do not support
each other's instruments. If a problem with operation should occur
which relates to the interface of the two manufacturer's equipment,
it is the responsibility of the owner to resolve the difficulty.
Several negative aspects must also be noted, not the least of
which is the marginal compatibility of the components. Because of
the small vacuum pumping capacity of the Hewlett Packard 5970 MSD,
the splitless sample introduction presents a major problem.
Collection parameters such as flow rate and duration can be
modified to insure no sample overload to the system. Ambient
sampling during high humidity conditions introduces water with each
sample. This presents problems in that water causes a dramatic
drop in sensitivity in the MSD.
There are few users of the complete system, thus flaws in the
design and construction (particularly of the thermal desorber) are
still being documented and corrected. Modifications are necessary
before the system can be used, an inconvenience which ultimately
results in more reliable numbers. It is anticipated that future
modifications will further improve performance.
Further recommendations include the use of a shorter, fused
silica transfer line rather than the nickle transfer line, and
perhaps 1/4" traps to reduce the amount of water added to the
system. These changes have not been implemented as of yet in this
lab, but are soon to come.
288
-------
I'irt fi''0r J ^. 30 t •' irfj.£-j ;~-
-------
Table I. Target Compounds
Calibration Correlation
Range, ng Coefficient , ^B
METHYLENE CHLORIDE 8.3 -
HEXANE 4.1 -
CHLOROFORM 9.3 -
1,1,1-JIRICHLOROETHANE 8.4 -
BENZENE 5.5 -
CARBON TETRACHLORIDE 10.0 -
TRICHLOROETHYLENE 9.2 -
TOLUENE 5.4 -
TETRACHLOROETHYLENE 10.1 -
ETHYLBENZENE 5.4 -
M,^XYLENE 10.8 -
STYRENE 11.3 -
0-XYLENE 5.5 -
(
T
i A
t .'
414.6 0.991
206.4 0.996
463.5 0.989
418.5 0.993
273.9 0.994
498.2 0.992
457.6 0.990
270.9 0.990
507.1 0.946
271.0 0.984
539.2 0.981
283.2 0.986
275.1 0.962
i ace | ,,o'
1. JC7
a.ac.f- y
? 3C '"*
i / s.ar.'-j y
\ . JCS -I - 4 QC^^
' 1* ' ! 4 ,"''
i - .- ' s • Bc? 1 X'
| 1--I -f^'. 3i4
1 U Prtt
--' J^S- "° , , : -^- £,
a loo saa 100 *B° __-J---^*-'
i •'c'-.vLtue i-ii-i^iac i
i . ...I
I ICC
i.aca-
|j
Figure 2. Calibration curves for representative compounds-
290
-------
OF AN ION-TRAP/MASS SPECTROMETER
ANALYSIS OF AMBIENT ORGANIC EMISSIONS
• Orth- David Haile, and Fred D. Hileman
St. T~ ? Co-> Environmental Sciences
^°uis, MO 63167
Hich
fintiis Wet>er-Grabau and Paul Kelley
?5S iR? Mat Corp.
r Oaks Parkway
. CA 95134-1991
analysis of ambient air for trace levels of organic
the field requires instrumentation which is
1*Plicn^a^ed in bein$ sensitive and versatile but is not
6 devi in °Peration. This, of course, is not fulfilled by any
ts°e °r raetnod °^ analysis. The ion trap/mass spectrometer
8^^^ ^n ^ts research and development stage appeared to at
some of tne desirable points mentioned. Therefore, an
tlle ion traP detector for detection limits using a
Probe was undertaken for acrolein in air. The
a tounH Was able to detect acrolein in a high hydrocarbon
de °Iein? Usi^g a selective chemical ionization method. The
acrylonitrile, styrene, and ethyl benzene at
. Tft5 from 24"° PPm for butene to 30° PPra f°r ethyl
of ion-trap detector was able to determine each component
sioh ?a^S sPectroraetry/mass spectroraetry methods which yield a
nent ^duced fragmentation spectrum characteristic of each
. tnougn the device is not commercially available for
analysis these results suggest that the instrument could
the needs stated at the beginning.
291
-------
INTRODUCTION
When carrying out ambient air analysis for organic compounds in
the field, whether at a waste site or an industrial plant site,
there is a need to have a detection device which is sophisticated in
being sensitive and versatile and yet easy to use and maintain. In
addition, a device should provide specificity with minimal handling
of the air sample. With all the above attributes the device should
be transportable without .the need for a sophisticated transportation
system. The use of such a device would allow for an analysis of
ambient air with a fast response so that evaluation of a site could
be accomplished with minutes turn around time. One instrument which
meets many of the attributes is the TAGA 6000 mobile MS/MS
laboratory(l-2).
Another instrument which is not commercially available as an
ambient air analysis device but does have many of the aforementioned
attributes utilizes ion trap technology(3-4). This technology, when
coupled with gas chromatography, has shown great promise in
defection limits and dynamic range for analysis. Recent advances in
the technology allows the device to be used as a selected ion trap
that traps only a specific ion from a mixture of ions. This suggests
the device would allow low detection limits in a complex mixture.
Another advance is that this device can analyze the selected ion
further by use of MS/MS techniques(4). The MS/MS techniques include
collision induced dissociation of the selected ion resulting in the
formation of collision induced ion fragments which are
characteristic of the ion and can be used to identify a component
present in a mixture. The second MS/MS method would be ion molecule
reactions where the reaction products arise from a specific
reaction.
Because of the potential of this instrument in ambient air
analysis, an evaluation of the device was undertaken. The device was
not designed as ambient air analyzer so some limitations on the
interface were necessary. The evaluation included the determination
of the detection limits of the ion trap/mass spectrometer device f°r
acrolein in an air stream. The analysis was to be done in such a way
that the air was to be sampled directly. Acrolein was examined in &
high hydrocarbon background which gives direct interferences in the
m/z range for acrolein at m/z 56. This allowed for a test of the
specificity of the device and a determination of the detection
limits with high chemical noise. The second evaluation included an
air bag sample which contained various components ranging in
concentrations from 300ppm to 2400ppm.
EXPERIMENTAL
Figure 1 shows a schematic cross section of the ion trap
detector/mass spectrometer as utilized in this evaluation. A
description of the instrument can be found in reference 5. The
ionization mode is selected to provide the best specificity in the
matrix being analyzed. The modes of ionization include chemical
ionization and electron impact.
The acrolein in air sample is prepared using a permeation
device which sets the level of the acrolein in the air stream
according to the temperature and flow rate. The air stream is
sampled through a split interface with a leak valve, VI, to control
the leak rate into the ITD/MS device. The static sample pressure i°
the device was generally at 2x10"5 Torr while the total pressure i°
292
-------
was jev*ce was in the 10" 3 Torr range. The difference in pressure
iTD/w6 to *"ne addition of He which improves the performance of the
anal s*nce the instrument has not been utilized for this type of
ysis previously the flow rate was not maximized nor was it
Drained if the instrument would perform satisfactorily if the
°^ 10 Torr in the device was due to the air being
In addition, the permeation device and the interface had a
"^n time constant which caused the analysis not to be
t at*taneous at this point. The combined time constant of the
e ace and the permeation device was 10 's of minutes for this
arigement while the interface response was one to two minutes.
e SraD bag sample contained butene(2400 ppm) , butadiene (1400
b;> acrylonitrile(690 ppm), styrene(350 ppm), and ethyl
ana?etle(320 ppm). The bag sample was analyzed by GC/FID prior to the
^tt ST"S *Y the ITD/MS device . The air in the bag was again sampled
£ y through the split interface and the ionization mode for the
sample was electron impact.
RESULTS AND DISCUSSION
air containing acrolein was admitted directly into the
ice without any prior separation. The majority of the
^2 and 02- Since the ion trap can be thought of as a cup
n ions from the sample, the predominate ions contained in
(cuP) would be those arising from N2 and Oj . This makes it
to ^tect the ppm level constituents of interest. To
the *P*-sh the detection at these trace levels the device can remove
°ns that arise from No and Oo or any selection of ions which
n°t of interest.
r nterferences which originate at the trace level such as
Sn?rbons can be minimized by other techniques in addition to the
ction mode mentioned above , In the determination of the
n limit for acrolein, C3H4.0, hydrocarbons were constantly
in the ITD/MS trap from a previous analysis involving crude
i°ns with m/z of 56 and 57 would interfere strongly with
lon °^ PPra levels °^ acrolein. To solve this problem the
C3H40 ....... > C^U^OH^ + CH^OH (1)
to selectively ionize the acrolein while minimal
a,rb°n ionization occurs. This is a soft ionization method
gli6 ^*ts in very little loss of signal from m/z 57 due to
^e lott Ure ^a snows the result of combining the use of the trap in
K the Selection mode, i.e. selecting only m/z 56-59 to be stored
;^e Ug traP and using reaction (1) to selectively ionize acrolein.
a eutif ? chemical ionization requires an alternative method for
ii °le? n^ that indeed the ion at m/z 57 is due to the protonated
n 6 of SA The ITD/MS is capable of accomplishing this task by the
<*llisi /MS methods (5, 7). These methods include the use of a
ii taci°n to ^raSraent the m/z 57 to yield fragment ions which are
a s se?ristic °^ tne protonated acrolein. The other MS/MS method
kitolei *on m°lecule reactions. In the case of protonated
f etio t*le reverse of reaction (1) can occur by increasing the
a1 the energy of the m/z 57 ion. The result is shown in Figure 2b
fc obs ' where the ion is excited selectively and results in
be d Ved ra//z ^3 due to Pr°tonated methanol. The m/z ions at 47
to a reaction product or protonation of ethanol which
293
-------
could be present as a trace contminant in the methanol. Further work
would need to be undertaken to be sure of the origin of this ion.
This reaction is confirmed by using pure acrolein to verify that the
reaction gives the observed 33 and 47 ions. The m/z 57 can also be
collisionally fragmented to yield characteristic peaks at m/z 29 and
27. This was accomplished using pure acrolein and exciting the
acrolein in He. When methanol is present, only the reverse of
reaction (1) is observed. Ions observed at 57 due to hydrocarbons
did not appear to protonate the methanol.
The detection limit was based on the signal to noise as
established by continuously monitoring m/z 57 at 50 ppm and 30 ppm
levels. The detection limit was determined to be 6 ppm. This is a
conservative detection limit since the interface for sampling was
not maximized. In addition, the ion trap could be filled with m/z 57
ions while ejecting unwanted ions. This filling of the trap was not
carried out in these experiments. Instead, the m/z 57 ions were
formed through reaction (1) for a 2 msec period followed by ejection
of unwanted ions and then the 57 ions were detected by ejecting them
out of the trap. Thus the trap was nearly empty and could easily
have contained many more ions. The detection limit based on the
observed signal would be expected to improve to sub ppm levels with
the increased number of ions in the trap,
The second evaluation involved the analysis of an air bag
sample containing, butene, butadiene, acrylonitrile, styrene, and
ethyl benzene. This sample was also directly analyzed and the ions
arising from air were again selectively removed by ejecting all the
ions below m/z 40 from the ion trap. As can be seen in Figures 3 the
mass spectrum obtained in electron impact ionization mode contains
ions which would be considered representative of the compounds in
the air bag sample. For example, the ion at 56 could be due to
butene while the ion at 106 could represent ethyl benzene. The
specificity however must be obtained by use of MS/MS techniques.
This was done using collisional induced dissociation for each m/z
ion which would represent the component in the bag. The resulting
CID spectrum should represent the component in the air bag sample.
Figure 4a shows the mass spectrum that results when m/z ions 106-10°
have been selectively retained in the trap while the other ions
shown in Figure 3 are removed. This is followed by kinetically
exciting the 106 ions and colliding them with He in the trap to
yield the fragment ions shown in Figure 4b. The fragments are
typical for a €2 benzene molecule as can be verified by obtaining a
CID spectrum for a standard of ethyl benzene. The butene also showed
a CID fragment spectrum which is typical of a C^Ug ion.
CONCLUSIONS
The ion trap detector/mass spectrometer appears from this
evaluation to have distinct promise as an analytical tool for the
determination of trace organics in ambient air. The simple interface
which samples air directly was shown to have a detection limit in &
high hydrocarbon background at the 6ppm level. This limit is
conservative and probably will be extended with improvements in the
ITD/MS as well as improving the interface utilized in this
evaluation. The device was able to identify and confirm the f
identification of a five component mixture in air with the levels 01
the components ranging from 2400 ppm to 300 ppm. The small size of _
the device itself indicates that it would be easily transportable i£
a suitable vacuum system was available. Although real time analysis
was not realized with this study it should be possible with a
properly designed interface.
294
-------
REFERENCES
Shushan, G, Debrou, Proceedings of the 1987 EPA/APACA
sitim Qn Measurements of Toxic and Related Air Pollutants .
Page 218. ~
2
B>B{. French> B-A- Thomson, W.R. Davidson, N.M. Reid, J.A.
Uckley, Mass Spectrometry in Environmental Sciences, Eds.,
£•". Karasek, 0. Hutzinger and S. H. Safe, Plenium, 1984, pp
' E- Fischer, Zeitschift Fur Phvsik. 1959, 156, 1-26.
V o o
j-jj. Stafford, Jr., P.E. Kelley, J.E.P. Syka, W.E. Reynolds,
loo/'^1 Todd, Int. Journ. Mass Spectrom. & Ion Processes. 60,
iV84, 85-98.
' 'I-N. Louris, R.G. Cooks, J.E.P. Syka, P.E. Kelley, G.C. Stafford,
r-> J.F.J. Todd, Anal. Chem.. 1987, 59(13), 1677-85.
6 h
^. Weber-Grabau, P.E. Kelley, J.E.P. Syka, S.C. Bradshaw,
ffMeQted at the 35th Conference on Mass Spectrometry and
^Ui^d Topics. May 24-29, Denver Ca., 1987, pp 1114.
1.
Ed-. Tandem Mass Spectrometry. J. Wiley & Sons,
295
-------
EXPERIMENTAL ARRANGEMENT
PERMEATION
DEV1CE 1TD/MS
Figure!. Schematic of interface for air sampling
Ion Trap Detector.
296
-------
INI
a) Selective detection of acrolein(50ppm)
57
33
IHt
'I""!""!""!"..!.
33 40 59
m/z
b) MS/MS 57 ion
60
59
70
40
50
m/z
Fi
gure 2. a) Detection of acrolein in air (50 ppm) using
selective trapping and methanol chemical ionization.
b) The selective ion molecule reaction of protonated
acrolein (m/z 57) with methanol.
297
-------
1007.
INT.
40
1
54
44 |
48
i..ll
6
•
77
67
0
i •
81 104 x67
93
87
ii
111
» !!! 135
I ii ii
, • i • . • l • . • 1 - i
L6B 67 7? 93.. 10.4.111119 . 135
6
i '
, , -,-j -^, -, | r- , . , . ,
0 80 180 120 140
m/z
Figure 3. The mass spectrum of the air bag sample
obtained while selectively removing m/z
below 40.
18 T
9-
a ) ETHYL BEN7PNF 1 ^
ISOLATED M/Z 106 lil_
LT
]C
I TEKP
It 2
90
uS
50|
~TnTpTTmui|ii mim JIM MI MI |iu tii mpi im m | IM«I i IM jiMti IMI |i* MI i
50-> 58 . 60 78 88 90 100 110 120 130 -H50
m/z
18 -1
"
9-
•
0-
HUIT
bj COLLISION INDUCED DISSOCIATION 150Q
TEHP
I
JL A
4Trt^WTrt1!Tnyf)1<"rTti'n'[TTtijlTri7v.Tf iliip'i n ri iir|fi
LZBQ]
us
Ljoj
50-K 50 60 70 80 98 IBB 118 128 130 -H59
m/z
Figure 4. a) The use of the ion trap to isolate m/z 105 to
108 where 106 is believed to be due to ethyl
benzene, b) The collision induced dissociation
of 106 ion which gives fragments which represen
ethyl benzene ion at m/z 106.
298
-------
AND ANALYSIS OF TOXIC VOLATILE
POLLUTANTS IN AMBIENT AIR USING
AUTOMATIC CANISTER-BASED SAMPLER
w- Sweet
'heric Chemistry Section
ois State Water Survey
An
eqy automatic canister-based sampler capable of collecting up to six
*n , 1-al air samples was used to monitor urban and rural air as part of
toxics study in Illinois. Contamination of samples from the
train and uptake of ambient air components by the sampling train
C ev&luated and found to be within acceptable limits. Samples were
Chf0 trated cryogenically and volatile organics analyzed by capillary gas
hy using FID and BCD detectors. Eight toxic compounds were
in most samples. In general, the concentrations found are
* to those reported elsewhere.
recent years, there has been considerable interest in methods for
and analysis of volatile toxic organics. Several of these
e s are important cancer risk factors in urban ambient air.
att ' tnere are relatively few ambient data on these materials because
aSUt ^uaHty standards have been set for them and they are not usually
*"n state monitoring programs. As more air toxics legislation is
a*-r monitoring agencies will need to develop sampling and
Cfll capabilities for a wide range of volatile organic compounds.
Tk
- e te are two general methods for sampling volatile organics. Air
Wn t^irou6h a solid sorbent such as activated charcoal or Tenax
solvent- or heat-desorbed for analysis. This method has been
g Sed> but it suffers from problems with sample breakthrough and
s °tds ^°rmation . Over the past few years, an alternative method that
S°me °^ these problems-' has been developed using passivated
steel canisters. With this technique, whole air samples are
b 8fini ^n inert stainless steel containers and concentrated
nakthj.0*11^ in the laboratory before analysis. This method avoids
Up'ie U^ and artifact problems and has the advantage of permitting
van analytical runs using the same sample. The principal
Pti 6 of tne metnod is the potential for contamination from and
°n to the numerous surfaces over which the collected air sample
299
-------
As part of a project to measure toxic airborne chemicals
urban/industrial areas in Illinois, a canister-based sampler was us . s
collect ambient air samples. Toxic volatile organics in these samp \
were then analyzed using cryogenic preconcentration, and compared to 6
samples taken at the same locations. These preliminary monitoring r6SUhu
and the results of quality assurance experiments are reported in c
paper.
Experimental Methods
IfS *P$'
Samples were collected in Summa®- polished, 6-liter stainless st f
canisters using an automatic, computer -control led canister-based salBLv
capable of collecting up to 6 separate samples (SIS Inc. , Moscow. * ^
The canisters were cleaned by filling with clean humid air, heating to ^j
C for one hour and then flushing three times with zero air. After a $
filling with zero air, the canisters were allowed to stand overnig^ i
then checked chromatographically . Any canister showing contamination
level greater than 10% of typical ambient levels of target compounds
recleaned.
tb»
Cleaned canisters were evacuated to 150 mm Hg and connected to $
sampler shown schematically in Figure 1. Sampling flow rates were & ^
provide 7-10 liters of air during the sampling period. Collected s^^
were returned to the laboratory and pressurized to two atmospheres ,
zero air. They were analyzed as soon as possible but no later than * .fl(i
weeks after collection. Analysis was by the cryogenic preconcent11* ^
method described by McElroy et al.* The only modification made "aS ef,
use of liquid nitrogen instead of liquid argon or oxygen as the c^ $$
Between 0.1 and 2.0 liters of diluted sample were passed through th* ^
at a flow rate of approximately 200 ml/min. After trapping, samplefl )
injected onto the chromatographic column by heating the trap to 85° c *$
water bath while back- flushing with carrier gas using &. six-port Valc"
sampling valve.
A Hewlett Packard 5890A gas chromatograph with flame ionizatlon
and electron capture (BCD) detectors was used. The chromatographic
was a 30 m, .325 mm ID, fused silica capillary column with a 1.0 0*
film of DB-5 bonded liquid phase (J&W Scientific, Rancho Cordova,
The column effluent was split 10:1 between the FID and ECD detector5
a VSOS capillary outlet splitter (SGE Inc., Austin, TX) . The tempet*
program was from -35°C to 200°C at 8°C/min. Compound identificati0
by retention time and quantification by peak height. Det
temperatures were 275°C and helium was the carrier gas.
The analytical method was calibrated using an 8-compound std^ v
containing 100 ppb of benzene, toluene and m-xylene and 10 ppe r
chloroform, methyl chloroform, carbon tetrachlorlde, triehloroethyl*0 ^
tetrachloroethylene (Scott Specialty Gases, Plumsteadville, PA)
standard mixture was diluted with clean humidified air to give a
standard with final concentrations of 50 ppb for the aromatics *
for the chlorinated hydrocarbons. The diluted standard was then
in the same way as a typical air sample.
300
-------
*he canister-based sampler was cleaned by purging with clean humid
v"Ue applying heat to all valves and sampling lines. The sampler was
for contamination and sample adsorption by collecting
co-located samples using the canister-based sampler and a 30-
CQI\S(- neea^-e- The needle was placed through a septum and provided a
lu. ant flow rate of 25 ml/min. to an evacuated 6-liter canister for two
The sampler was programmed to sample at the same flow rate during
auie period. After collection, the canisters were brought to two
pressure with zero air and analyzed as described above.
ivd evaluate the analytical method, a sample containing 41 compounds
Systfi ln ambient air was obtained from U.S. EPA, Environmental Monitoring
e s laboratory. The eight compounds in the standard as well as four
Vere quantified in this sample. Calibration standards for the
comPounds were prepared using the static dilution bottle
The results are shown in Table I. With the exception of
all of the concentrations agreed within analytical error,
Th
6 canister-based sampler was then evaluated under normal operating
°ns at 28° C. The analytical results for simultaneous samples of
<tkient air collected by the sampler and the needle orifice are
Table ll- After an initial cleaning, high levels of
oroethyl«ne were found in sampler-collected air. The sampler was
again and retested. In the second test, all eight standard
were above detection limits and their concentrations were the
kin analytical error in both needle- and sampler-collected
Some reductions in the concentrations of compounds with
t^nies longer than that of m-xylene were noted in the samples
by the automatic sampler.
monitoring results for eight volatile organic chemicals
Table III, Samples were taken in two major urban areas--
' Louis and Chicago -- and a rural area in Illinois. The urban
in residential- areas near a wide variety of industrial sources.
site is located 50 km downwind from the nearest urban area and
ive of rural conditions in the Hidwest, Average
Sa of the target chemicals in midday grab samples were similar
of ^eP°rted elsewhere.6 The urban values had large fluctuations for
compounds. Minimum urban concentrations were similar to rural
maximurn concentrations were as much as 100 times greater than
values. The canister -based sampler was able to collect
time' integrated samples. This allows a better estimation of
jy ient concentrations and better detection of short-term releases
chemicals.
(^ et\ty, ^ a recently developed canister-based automatic air sampler,
^t^^tlons of at least eight toxic volatile organic chemicals could be
u^d \~ in ambient urban air samples. Further development of this
3ft lilt- ° Jua^e it possible to measure many more compounds. The
*itil^ °^ collecting time -aver aged samples is important for
ng average concentrations and detection of short term releases.
301
-------
References
1. V. E. Thomson, A. Jones, E. Haemisegger, B. Steigerwald, '"
Toxics Problem in the United States: An Analysts of Cancer
Posed by Selected Air Pollutants," J. Air Pollut. Cont.
535. (1985).
2. J.F. Walling, "The Utility of Distributed Air Volume Sets *
Sampling Ambient Air using Solid Absorbents," Atmos. Environ. IS
(1984) .
. ^
3. W.A. McClenny, J.D. Pleil, T.A. Lumpkin, K.D. Oliver, "UpdaC/A
Canister-Based Samplers for VOC's," Proceedings of tfoe
Symposium on Measurejtgivt of Toxic Air Pollutants. Hay ^987 • Rft
NC. APCA Publication VIP-8:253, Air Pollution Control Associ
Pittsburgh, PA.
cell*'
4. F.F. McElroy, V.L, Thompson, D.M. Holland, W.A. Lonneman, R.L- 9 £ 6t
"Cryogenic Preconcentration - Direct FID Method for Measureme11
Ambient NMOC: Refinement and Comparison with GC Speciation,
Pollut. Cont. Assc. 36 710. (1986).
C. Morris, R. Berkley, J. Bumgarner, "Preparation of Multicoropn jj
Volatile Organic Standards Using Dilution Bottles,"
1585. (1983).
6. R. Brodzinsky, H.B. Singh, "Volatile Organic Chemicals *
Atmosphere: An Assessment of Available Data," Report No. EPA*
93-027a Environmental Sciences Research Laboratory, U.S-
Research Triangle Park, NC. 1983.
302
-------
TABLE II. Audit of Cannister-Based Sampler
Compound
a ECD detector
b FID detector
Retention Time
(roln)
Concentration
Needle Safflp
lel
chloroform3
methylchloroforma
benzene
carbon tetrachloride3
trichloroethylenea
toluene""1
tetrachloroethylenea
m-xylene^
4.4
5.0
5.4
5.5
6.8
9.1
10.5
12.3
0.10
0.50
0.14
0.14
0.10
0.32
0.04
0.14
•
o.ii
£ 1
0.51
~ 1
,»? 5>
^midday samples, n-10
12-Hour integrated samples. DAY - 6AM-6PM, NIGHT - 6PM
ND - majority of samples below detection limit
. 6AM
304
-------
FOR T°XICS INTERFACE AND ANALYTICAL SYSTEMS
* AMBIENT AIR SAMPLES
p °ayton
lce
McA1 lister
F' JonS
M°ore
Triangle Park, North Carolina
j ll-tldetector gas chroraatographic (GC/MD) analytical system,
l"easmr-r :^uded a sample interface, was developed and tested for
SainPle ng tlle concentration of 36 toxic compounds in ambient air
ait|hient' The sainple interface system can deliver equal volumes of
V*CuUis Lit samPle frora 6- liter canisters under pressure or under
rfai 8ny analytical device without drying the sample. The sample
sam °an inJect t'ie sample directly, cryogenically preconcentrate
e- or cause the sample to adsorb onto an adsorbent.
* 0 •; ^ 3700 gas chromatograph column was configured with a
^ficislo] 5 mm, DB-624 Megabore fused silica capillary, followed by a
• ctto r Sp-'-it;t:er which routed one- tenth of the gas through an
°apture detector (ECD) , and the remainder through a photo-
. ? letector (FID) in series with a flame ionization detector
r Qf i-hour pressurized ambient air samples obtained during the
Sa"iplc fr°m 15 urban sites, were analyzed on the GC/MD system.
°al pr •rom each site was analyzed twice to determine the analy-
t- !i°n; one duplicate sample from each site was analyzed to
e Sflnipling and analysis precision. Two samples from each
t, ana^yzed by gas chromatography mass spectrometry (GC/MS) to
• accuracy of compound identification.
P ,
ion £ 1 air toxics compounds were identified ranging in concen-
' Ppbv ?m ^ust above the detection limit (0.2 to 0.9 ppbv) to
syst °r methylene chloride. Compound identification by the
1 was confirmed by GC/MS in 91.4% of the cases.
v *"6ci.sj
* ^ for *n terras °f absolute percent differences averaged from
•ePlicate analyses and duplicate sample analyses.
305
-------
INTRODUCTION
Radian began the study of air toxics compounds in ambient air
samples collected for the Nonmethane Organic Compound (NMOC) Mon-
itoring Program in 1986. For the 1987 monitoring program, NMOC
ambient air samples were collected at 32 sites from 6:00 AM to 9:00 A#
at about 15 psig in 6-liter (L) stainless steel canisters. From 15 of-
the NMOC sites, after NMOC analysis, the samples were also analyzed on
the GC/MD system for the compounds listed in Table 1. In a related
study, Radian began collecting and analyzing 24-hour ambient air
samples at 19 sites. The latter samples are analyzed for air toxic
compound (Table 1) concentrations. The 24-hour samples will continue
to be collected and analyzed through September 1988.
Delivery of the ambient air samples from the canisters to the
GC/KD. analytical system involved transferal of the sample from the
canister under 0.0 to 10.0 psig pressure for the 3-hour samples, and
under 0.5 to 14.0 inches Hg vacuum in the case of the 24-hour ambient
air samples. Radian Corporation developed a sample interface system
which reliably delivers a constant volume of cryogenically preconcen-
trated ambient air to either the GC/MD or GC/MS without drying the
sample.
The system interface was developed considering contamination,
memory effects, and repeatability. The potential for contamination °*
the ambient air samples was minimized by the design and selection of
materials of construction for the interface. The connecting tubing °*
the interface was 1/8-inch and 1/16-inch o.d. chromatographic-grade
stainless steel. Each routing valve, shutoff valve, and fitting was
made of 316 stainless steel. The preconcentration trap assemblies
were constructed of chromatographic-grade stainless steel and fiHed
with 60/80 mesh glass beads. Daily baseline checks.consisted of
sampling and analyzing humidified zero-grade air immediately after
calibration. Zero air analyses have shown the interface system to
remain essentially contamination free.
The elimination of compound memory from previous samples was
achieved by the use of heat-traced temperature controlled components-
Each part of the interface which contacted the sample before and afte
the analytical trap was temperature controlled. The 8-port gas
sampling valve was enclosed in an oven and maintained at 160 C with a
active temperature controller. Transfer lines were maintained at
160 C with a separate active temperature controller. The analytic*1
sample traps were heated to over 200°C during the thermal desorpti°n
cycle to remove any residual compounds.
Repeatability has been accomplished by using a high resoluti°n
pressure/vacuum gauge to measure sample loading pressures accurately-
Trapping temperature was set at -185 C, and thermal desorption was
accomplished by using He carrier gas and a ramp-and-soak temperature
controller capable of delivering 25 amperes of electrical current at
120 volts AC to a 1000-watt heater inbedded in a brass block
306
-------
ai-ning the sample trap. Repeated measurements of calibration
rds indicated that the samples can be delivered at a constant
e with a precision of less than 6 percent.
R
"• VARIAN 3700 gas chromatograph, configured with a photo-
i-on detector (PID) , a flame ionization detector (FID) , and an
30^Lr°n caPture detector (ECD) performed the air toxics analyses. A
h eter DB-624 chromatographic column separated the target compounds
n§ retention times greater than that of methylene chloride.
RESULTS
foj. Estimated detection limits for GC/MD and GC/MS analytical systems
Ijj T arget compounds in the Urban Air Toxics Program (UTAP) are given
toer ^' Four pairs of compounds coelute on the DB-624 chroma-
bg-. P^c column and cannot be separated by the GC/MD system --
H.j, 1rie/li 2-dichloroethane ; n-octane/cis-1, 3-dichloropropylene;
ejtj, ene/p-xylene; and styrene/o-xylene. All of these pairs with the
are J °n of m-xylene and p-xylene which have the same mass spectra
identified on the GC/MS.
at e en 3-hour ambient air samples were taken for air toxics analyses
9'fift °f 15 urban sites. The samples were taken from 6:00 AM to
•UIJ AU . r
IS r. . ln evacuated stainless steel canisters and were at 10 to
?it 6 at the end of the sampling period. The samples were analyzed
flam their NMOC content by the cryogenic preconcentration and
t$j. *°nization detection (PDFID) method, and then for air toxics
also c°mpounds by the GC/MD system. Two samples from each site were
analyzed by GC/MS as confirmations of the GC/MD analyses.
*l\ M, 2 summarizes the air toxics compound identifications for
"t p ,6 3-hour ambient air samples. Eighteen of the target compounds,
by ti rs °f compounds (for those that coelute) are listed in Table 2
frequency of identification (number of cases), minimum,
~m, and mean concentrations in parts per billion by volume
-j m, ' The most frequently identified compound was toluene, followed
tira^.^"xylene, styrene/o-xylene, and 1,1,1- trichloroethane . Concen-
chlorljS ranged from the detection limit to 94.8 ppbv for methylene
^ t e' The data in Table 2 do not include the identification from
ePlicate analyses.
CoitlPai-i 6 ^ summarizes the GC/MD and GC/MS compound identification
lsons into four kinds:
Positive GC/MD - Positive GC/MS;
Positive GC/MD - Negative GC/MS;
Negative GC/MD - Positive GC/MS; and
Negative GC/MD - Negative GC/MS.
ir> i»hi fre were 156 cases Positive GC/MD - Positive GC/MS comparisons
% GC/M a ComP°und identified by the GC/MD system was confirmed by
C/*,,, /MS analytical system. The comparisons classified as Negative
Negative GC/MS were also considered as positive confirmations.
307
-------
The total positive comparisons expressed in percentage was 91.5.
Incorrect comparisons included those compounds that were identified b
one of the analytical systems, but not by the other. They include
those compounds identified by the GC/MD system but not confirmed by
the GC/MS and those compounds reported by GC/MS but not identified by
the GC/MD system.
QUALITY ASSURANCE
Daily calibrations are made using in-house standards of 36
compounds at 50 ppbv. Care had to be taken in calculating the concert*
trations of o-, m- , and p-xylene and toluene because the chloropretie
"standard" contained the xylenes as stabilizing agents and toluene &s
a contaminant. In-house standards were made in a 33-L Summa treated
canister at 10 ppmv of each of 36 air toxics compounds in zero grade
air that had been humidified with HPLC-grade water. Samples from the
33-L canister were used to make up 10 ppbv in-house standards in 6-L
stainless steel canisters for use in the daily calibrations of the
GC/MD system. After each daily calibration, response factors were
calculated for each compound so long as the daily response factors
each compound was within + 20% of the average calibration factor
previous calibration data.
Reported concentration was quantified using the response of the
FID detector. Compound identification was done using the retention
time and the ratios of area counts on whichever of the detectors
responded to the compounds. Depending on how many, and which de-
tectors responded, the compound identifications were classified as *
high
-------
, lower part of Table 4 summarizes statistics for duplicate
°Ur ambient air samples. Sampling and analysis precisions are
» ven in terms of % CV and Absolute % Differences for the
ence"level pairs shown. As seen in the replicate pairs dis-
a^ove> an inverse relationship is seen between confidence level
3nd Precision, e.g. H-H (6.1% CV, 8.6 Absolute % Difference) and
obt -6% cv> 13.6 Absolute % Difference). The same relationship
(16 nS between H"H (fi-1% cv- 8-6 Absolute % Difference) and H-M
tnv ^' 22.9 Absolute % Difference). However, the precisions
tiot° g tne low compound identification level, L (H-L, and M-L) do
aPpear to follow the same pattern. On the other hand, the number
th 9ses f°r tne H-L and M-L comparisons is only four cases each; and
Po °re» it is felt that these data are not representative of a
Cotif trend. Additional data for these cases would be necessary to
m °r ^eny c^e hypothesis that the precision is inversely pro-
Port-* eny e ypoess a e precson s nversey pro
pt , °nal to the confidence level (or directly proportional to the
alDility of making an error in compound identification) .
ana, "ecause the duplicate precision includes both sampling and
pt. ygis variability (expressed as precision) and the replicate
^UD! * °n inv°lves only the analytical error, one would expect the
tej Cate precision to be larger than the replicate precision. This
V j 0nshiP holds for H-M and M-M comparisons (see Table 6 and 7),
H0tll Oes not. hold for H-H and H-L comparisons. These results will be
tt^ °rsd as more 3 -hour data becomes available to be able to dis-
ish between random behavior and actual differences.
309
-------
TABLE 1. ESTIMATED DETECTION LIMITS FOR AIR TOXICS COMPOUNDS
__-_^ — ,
Compound
Methylene chloride
trans -1,2- Dichloroethylene
1, 1-Dichloroe thane
Chloroprene
Bromochlorome thane
Chloroform
1,1,1 -Trichloroethane
Carbon tetrachloride ,
Benzene/1 , 2 -Dichloroe thane
Benzene
Trichloroethylene
1 , 2-Dichloropropane
Bromodichlorome thane
trans -1 , 3-Dichloropropylene
Toluene
n-Octane/c-1, 3-Dichloropropylene
cis-1 , 3-Dichloropropylene
1,1,2- Tr ichloroe thane
Tetrachloroethylene
Dibromochlorome thane
Chlorobenzene
m/p-Xylene
Styrene/o-Xylene
o-Xylene
Bromoform
1,1,2, 2-Tetrachloroethane
m-Dichlorobenzene
p-Dichlorobenzene
o-Dichlorobenaene
GC/MDh
ppbv
0.4
0.7
1.3
2.6
1.1
0.9
0.4
0.6
2.2
0.5
0.9
0.6
1.4
0.6
0.5
0.4
1.0
0.8
0.8
1.7
0.4
1.2
0.2
0.2
0.07
0.03
0.3
0.3
0.4
ppbv
--'
0.5
- 0
0-3
,.. 3
(M
A
2-°
\
O-3
A
O-3
,.
n 1
U>
n i
o.
o.*
O-*1
o 5
UF
0 5
V'
A, 6
n ^
V •
o)
v *
o'*
V
o.3
^
V '
0-2
0.'2
0,2
— — — — — 1___ — —
The following compounds were not resolved on DB-624 analytical
acetylene, 1,3-butadiene, vinyl chloride, chloroethane,
, propylene, and bromomethane
Benzene and 1,2-dichloroethane coelute on DB-624 column
^Quantitated by FID
n-OcCane and cis-1,3-dichloropropylene coelute on DB-624 column
^Quantitated by ECD
m-Xylene and p-xylene coelute on DB-624 column
^Styrene and o-xylene coelute on DB-624 column
The GC/MD interface system samples about 250 mL of air (corrected
.atmospheric pressure)
The GC/MS interface system samples about 500 mL of air (corrected
atmospheric pressure)
310